Meeting record processing method and device

Abstract

The present invention provides a method and device for processing meeting records, particularly relating to the field of big data. The method comprises: determining, based on the current meeting record, a similar record with the highest similarity from historical meeting records; determining the optimal person to handle the current meeting record based on the similar records; determining, based on the source of the current meeting record, a set of current meeting records corresponding to different sources; obtaining a comprehensive weight of the current meeting record based on the weight of the host corresponding to the current meeting record in the current meeting record set; determining, based on the correspondence between the current meeting record and the optimal person to handle the record, the record to be processed corresponding to the optimal person to handle the record; and sending the record to be processed corresponding to the optimal person to handle the record and the corresponding comprehensive weight to the optimal person to handle the record to be processed based on the comprehensive weight. The present invention can improve the speed and accuracy of meeting record processing, thereby improving the efficiency of meeting record processing.

Description

Technical Field

[0001] The present invention relates to the field of conference analysis technology, particularly to the field of big data, and more particularly to a conference record processing method and device. Background Art

[0002] In various units and companies, after a meeting, the relevant meeting minutes need to be processed and analyzed to clarify the main points and key points of the meeting content, thereby providing a basis for subsequent decision-making and business processing based on the meeting results. Different meeting minutes need to be assigned to corresponding meeting minute processing personnel for processing. However, in the existing technology, the nature of each meeting minute needs to be manually analyzed and studied to determine the processing personnel to be assigned to the meeting minute. In addition, when a processing personnel is assigned to multiple meeting minutes, the processing personnel also needs to continue to analyze and study the nature of the multiple assigned meeting minutes to determine the priority of different meeting minutes, so as to subsequently process the different meeting minutes in the corresponding order and processing method based on the priority.

[0003] However, the above-mentioned method of assigning meeting minutes to processing personnel is time-consuming due to its reliance on manual labor, resulting in a slow speed in the allocation of meeting minutes. Moreover, the accuracy of the meeting minute allocation also depends on the ability and work experience of the relevant allocation personnel, and the accuracy of the allocation is low. During the allocation process, it is very likely that the meeting minutes will be assigned to a processing personnel who is not suitable for processing the meeting minutes, which will cause the subsequent processing by the processing personnel to be time-consuming and prone to errors. Therefore, the above-mentioned method of assigning meeting minutes to processing personnel will result in a poor speed and accuracy in the overall processing of meeting minutes. The above-mentioned method of determining the priority of meeting minutes is also dependent on manual labor, and is therefore also time-consuming, resulting in a slow speed in determining the priority of meeting minutes. Moreover, the accuracy of determining the priority of meeting minutes also depends on the ability and work experience of the relevant processing personnel, and the accuracy of determining the priority is low, which may cause the processing personnel to process the meeting minutes in the wrong order and manner, resulting in poor processing results. Therefore, the above-mentioned method of determining the priority of meeting minutes will result in a poor speed and accuracy in the overall processing of meeting minutes.

[0004] In summary, the existing technology has the problem that the processing speed of conference records is slow and the accuracy is low, thereby making the processing efficiency of conference records low. Summary of the Invention

[0005] One object of the present invention is to provide a method for processing meeting records to address the problem of slow processing speed and low accuracy in the prior art, which results in low efficiency in meeting record processing. Another object of the present invention is to provide a meeting record processing device. Another object of the present invention is to provide a computer device. Yet another object of the present invention is to provide a readable medium.

[0006] In order to achieve the above objectives, one aspect of the present invention discloses a method for processing conference records, the method comprising:

[0007] According to the current meeting record, determine the similar record with the highest similarity from multiple historical meeting records; based on the similar records, determine the optimal person to handle the current meeting record;

[0008] Determining, based on the sources of multiple current meeting records, current meeting record sets corresponding to different sources; and obtaining a comprehensive weight of all current meeting records based on the weights of the moderators corresponding to the current meeting records in the current meeting record set;

[0009] According to the correspondence between the current meeting record and the optimal processing personnel, the pending records corresponding to each optimal processing personnel are determined; the multiple pending records corresponding to the optimal processing personnel and the corresponding comprehensive weights are sent to the optimal processing personnel, so that the optimal processing personnel processes the corresponding pending records based on the comprehensive weights.

[0010] Optionally, further including:

[0011] Before determining the most similar record from multiple historical meeting records based on the current meeting record,

[0012] Performing voice-to-text processing on the preset current conference voice to obtain a current conference record corresponding to the voice source of the current conference voice;

[0013] A projection set is formed based on multiple preset current conference projection images, and text is extracted from one of the current conference projection images in the projection set to obtain the current conference record corresponding to the projection source of the projection set.

[0014] Optionally, forming a projection set based on multiple preset current conference projection images includes:

[0015] Based on the pixel information of all current conference projection images, the pixel similarity between all current conference projection images is obtained;

[0016] Repeating the step of forming a projection set until there is no current conference projection image that is not included in the projection set, the step of forming a projection set includes:

[0017] Select a target image from the current conference projection images that are not included in the projection set;

[0018] One of the projection sets is formed based on the selected images in the current conference projection images that are not included in the projection set, except the target image, whose pixel similarity with the target image is greater than a preset pixel similarity threshold, and the target image.

[0019] Optionally, forming a projection set based on multiple preset current conference projection images includes:

[0020] Extract text from all current conference projection images to obtain the corresponding text content;

[0021] Based on the text content, obtaining the text similarity between all corresponding current conference projection images;

[0022] Repeating the step of forming a projection set until there is no current conference projection image that is not included in the projection set, the step of forming a projection set includes:

[0023] Select a target image from the current conference projection images that are not included in the projection set;

[0024] One of the projection sets is formed based on the selected images in the current conference projection images that are not included in the projection set, except the target image, whose text similarity with the target image is greater than a preset text similarity threshold, and the target image.

[0025] Optionally, determining a similar record with the highest similarity from multiple historical meeting records based on the current meeting record includes:

[0026] Performing word segmentation processing on the current meeting record to obtain a corresponding current vocabulary, and performing word segmentation processing on the plurality of historical meeting records to obtain historical vocabulary corresponding to the historical meeting records;

[0027] determining a record similarity between the current meeting record and each of the historical meeting records based on the current vocabulary and the historical vocabulary corresponding to each of the historical meeting records;

[0028] The historical meeting record with the highest record similarity is determined as the similar record.

[0029] Optionally, determining the record similarity between the current meeting record and each of the historical meeting records based on the current vocabulary and the historical vocabulary corresponding to each of the historical meeting records includes:

[0030] Taking the intersection of the current vocabulary and the historical vocabulary corresponding to the historical meeting records to obtain a vocabulary intersection;

[0031] Taking a union of the current vocabulary and the historical vocabulary corresponding to the historical meeting records to obtain a vocabulary union;

[0032] Based on the number of words in the vocabulary intersection and the number of words in the vocabulary union, the record similarity between the current meeting record and the historical meeting record is obtained.

[0033] Optionally, determining the optimal person to handle the current meeting record based on the similar records includes:

[0034] Based on the similar records, corresponding historical processing information is obtained;

[0035] Determining, based on the historical processing information, a plurality of processing times corresponding to each historical processing person;

[0036] Based on the current time and the processing time, obtaining time differences corresponding to multiple processing times of each historical processing personnel, and determining a maximum time difference based on all time differences corresponding to all historical processing personnel;

[0037] Obtaining a processing weight corresponding to the historical processing personnel according to multiple time differences corresponding to the historical processing personnel, a maximum time difference, and a preset correction coefficient;

[0038] The historical processing personnel with the largest processing weight is determined as the optimal processing personnel.

[0039] Optionally, determining, based on the sources of the plurality of current meeting records, sets of current meeting records corresponding to different sources respectively includes:

[0040] The current conference record whose source is the screen projection source is determined as the current conference screen projection record, and the current conference record whose source is the voice source is determined as the current conference voice record;

[0041] Determining a first current conference record set corresponding to a screen projection source according to the current conference screen projection record;

[0042] A second current conference record set corresponding to the voice source is determined based on the current conference voice record.

[0043] Optionally, determining a first current conference record set corresponding to a screen projection source based on the current conference screen projection record includes:

[0044] Perform word segmentation on each of the current conference screen projection records to obtain corresponding screen projection vocabulary;

[0045] Determining a first current similarity between all current conference screen projection records based on the screen projection vocabulary corresponding to the current conference screen projection record;

[0046] Repeating the step of forming a first current conference record set until there is no current conference screen projection record that is not included in the first current conference record set, the step of forming the first current conference record set includes:

[0047] Selecting a first target record from the current conference projection records that are not included in the first current conference record set;

[0048] Based on the first candidate record and the first target record in the current conference projection records that are not included in the first current conference record set, the first current similarity between the other records except the first target record and the first target record is greater than the preset current similarity threshold, one of the first current conference record sets is formed.

[0049] Optionally, determining the first current similarity between all current conference screen projection records based on the screen projection vocabulary corresponding to the current conference screen projection record includes:

[0050] Intersections of two screen projection words corresponding to all the current conference screen projection records are taken to obtain a plurality of corresponding first intersections;

[0051] Taking the union of two screen projection words corresponding to all the current conference screen projection records to obtain multiple corresponding first unions;

[0052] Based on the number of words in the first intersection and the number of words in the first union, a first current similarity between the corresponding two current conference projection records is obtained.

[0053] Optionally, determining a second current conference record set corresponding to a voice source based on the current conference voice record includes:

[0054] Perform word segmentation processing on each of the current conference voice records to obtain corresponding voice vocabulary;

[0055] determining a second current similarity between all current conference voice records based on the voice vocabulary corresponding to the current conference voice record;

[0056] Repeating the step of forming a second current conference record set until there is no current conference voice record that is not included in the second current conference record set, the step of forming the second current conference record set includes:

[0057] Selecting a second target record from the current conference voice records that are not included in the second current conference record set;

[0058] One of the second current meeting record sets is formed based on the second candidate records and the second target record, whose second current similarity with the second target record in other records of the current meeting voice records other than the second target record is greater than the preset current similarity threshold.

[0059] Optionally, determining the second current similarity between all current conference voice records based on the voice vocabulary corresponding to the current conference voice record includes:

[0060] Intersections of two voice words corresponding to all the current conference voice records are taken to obtain a plurality of corresponding second intersections;

[0061] Taking unions of two voice words corresponding to all the voice records of the current conference to obtain a plurality of corresponding second unions;

[0062] Based on the number of words in the second intersection and the number of words in the second union, a second current similarity between the corresponding two current conference voice records is obtained.

[0063] Optionally, obtaining the comprehensive weight of all current meeting records based on the host weight corresponding to the current meeting record in the current meeting record set includes:

[0064] Determine whether the source corresponding to the current conference record set is a screen projection source or a voice source,

[0065] If it is a screen projection source, based on the current meeting record, obtain the number of projections of the current meeting projection image included in the corresponding projection set; based on the number of projections corresponding to the current meeting record and the host weight, obtain the corresponding sub-weight; superimpose the sub-weights corresponding to all current meeting records in the current meeting record set to obtain the projection sub-weight corresponding to the current meeting record set, and use the projection sub-weight as the projection sub-weight of each current meeting record in the current meeting record set;

[0066] If the source is voice, the host weights corresponding to all current meeting records in the current meeting record set are superimposed to obtain the voice sub-weight corresponding to the current meeting record set, and the voice sub-weight is used as the voice sub-weight of each current meeting record in the current meeting record set;

[0067] Based on the screen projection weights or voice weights corresponding to all current meeting records, the corresponding comprehensive weight is determined.

[0068] Optionally, determining the corresponding comprehensive weight based on the screen projection weights or voice weights corresponding to all current conference records includes:

[0069] Determine whether the current conference record is only a screen projection record of the current conference, only a voice record of the current conference, or both a screen projection record and a voice record of the current conference;

[0070] If only the current conference screen projection record is used, the corresponding comprehensive weight is determined based on the screen projection weight and the preset screen projection coefficient;

[0071] If only the current conference voice record is used, the corresponding comprehensive weight is determined based on the voice sub-weight and the preset voice coefficient;

[0072] If it is both a screen projection record and a voice record of the current meeting, the corresponding first comprehensive weight is determined based on the screen projection weight and the preset screen projection coefficient, and the corresponding second comprehensive weight is determined based on the voice weight and the preset voice coefficient. The first comprehensive weight and the second comprehensive weight are added together to obtain the comprehensive weight.

[0073] Optionally, determining the to-be-processed records corresponding to each optimal processing personnel according to the correspondence between the current meeting records and the optimal processing personnel includes:

[0074] Based on the correspondence between each current meeting record in the current meeting record set and the optimal processing personnel, determining the number of current meeting records in the current meeting record set corresponding to different optimal processing personnel, and taking the optimal processing personnel corresponding to the largest number of current meeting records as the optimal processing personnel corresponding to the current meeting record set;

[0075] Selecting a representative record from each current meeting record set, and determining a correspondence between the optimal processing person and the representative record based on a correspondence between the representative record and the current meeting record set and a correspondence between the current meeting record set and the optimal processing person;

[0076] According to the correspondence between the optimal processing personnel and the representative records, all the representative records corresponding to the optimal processing personnel are determined as the records to be processed corresponding to the optimal processing personnel.

[0077] Optionally, sending the multiple pending records corresponding to the optimal processing personnel and the corresponding comprehensive weights to the optimal processing personnel so that the optimal processing personnel processes the corresponding pending records based on the comprehensive weights includes:

[0078] After sorting the multiple pending records and the corresponding comprehensive weights corresponding to the optimal processing personnel according to the comprehensive weights, the records are sent to the optimal processing personnel so that the optimal processing personnel processes the corresponding pending records based on the sorted comprehensive weights.

[0079] In order to achieve the above objectives, another aspect of the present invention discloses a conference record processing device, the device comprising:

[0080] An optimal personnel determination module is used to determine a similar record with the highest similarity from multiple historical meeting records based on the current meeting record; and based on the similar records, determine the optimal person to handle the current meeting record;

[0081] A comprehensive weight determination module is used to determine, based on the sources of multiple current meeting records, a set of current meeting records corresponding to different sources; and obtain a comprehensive weight of all current meeting records based on the weights of the host corresponding to the current meeting records in the set of current meeting records;

[0082] A processing module is used to determine the pending records corresponding to each optimal processing personnel based on the correspondence between the current meeting records and the optimal processing personnel; and send multiple pending records corresponding to the optimal processing personnel and the corresponding comprehensive weights to the optimal processing personnel, so that the optimal processing personnel processes the corresponding pending records based on the comprehensive weights.

[0083] The present invention also discloses a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the method described above is implemented when the processor executes the program.

[0084] The present invention also discloses a computer-readable medium on which a computer program is stored. When the program is executed by a processor, the method described above is implemented.

[0085] The meeting record processing method and device provided by the present invention determine the similar record with the highest similarity from multiple historical meeting records based on the current meeting record; and determine the optimal processing personnel of the current meeting record based on the similar records. It can determine the similar record closest to the current meeting record from the actual historical meeting records based on the similarity that fully reflects the degree of similarity between the meeting records, and fully considers the rule that if a processing personnel has processed a historical meeting record, then when processing the current meeting record similar to the historical meeting record, the processing proficiency is higher. Therefore, the above steps can not only improve the accuracy of determining similar records, but also improve the accuracy of determining the optimal processing personnel. Moreover, the above steps are relatively simple, with low computational complexity, and can be automatically executed with low dependence on manual labor. Therefore, the above steps can not only improve the speed of determining similar records, but also improve the speed of determining the optimal processing personnel, so that the above steps can improve the accuracy and speed of the overall meeting record processing. By determining a current meeting record set corresponding to different sources based on the sources of multiple current meeting records; and obtaining a comprehensive weight of all current meeting records based on the host weight corresponding to the current meeting record in the current meeting record set, it can fully consider the impact of the source of the meeting record on the importance of the meeting record, and the rule that the importance of the host corresponding to the meeting record is positively correlated with the importance of the meeting record, and realize the comprehensive weight of the corresponding meeting record based on the source and the host weight that can measure the importance of the host, so that the determined comprehensive weight can accurately reflect the importance of the meeting record, and the importance is positively correlated with the priority, so the determined comprehensive weight can also accurately reflect the priority of the meeting record, so the above steps can improve the accuracy of the determined comprehensive weight, and the above steps can form a set, and can perform related operations and processing with the set as the processing unit, indirectly reducing the amount of data processed. In addition, the above steps can be automatically executed, with low dependence on manual labor, so it can also improve the speed of determining the comprehensive weight, so the above steps can improve the accuracy and speed of the overall meeting record processing.By determining the pending records corresponding to each optimal processing personnel based on the correspondence between the current meeting records and the optimal processing personnel; sending the multiple pending records corresponding to the optimal processing personnel and the corresponding comprehensive weights to the optimal processing personnel, so that the optimal processing personnel can process the corresponding pending records based on the comprehensive weights, the specific pending meeting records assigned to the specific processing personnel can be suitable for the processing of the processing personnel, which is conducive to enabling the processing personnel to skillfully process the assigned meeting records, and enabling the processing personnel to directly determine the order and method of processing the meeting records based on the accurate comprehensive weights, without the need for the processing personnel to conduct additional analysis of the meeting records, so that the accuracy and speed of the meeting record processing can be directly improved, and the step of assigning meeting records can also be automatically executed, reducing the degree of dependence on manual labor, so that the speed of meeting record processing can be further improved. In summary, the meeting record processing method and device provided by the present invention can improve the speed and accuracy of meeting record processing, thereby improving the efficiency of meeting record processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0087] Figure 1 A schematic diagram showing a flow chart of a method for processing conference records according to an embodiment of the present invention;

[0088] Figure 2 A schematic diagram showing an optional step of determining similar records according to an embodiment of the present invention is shown;

[0089] Figure 3 A schematic diagram showing an optional step of determining the optimal processing personnel according to an embodiment of the present invention is shown;

[0090] Figure 4 A schematic diagram showing an optional step of determining the records to be processed corresponding to each optimal processing person according to an embodiment of the present invention is shown;

[0091] Figure 5 A schematic diagram showing an optional step of sending multiple to-be-processed records and corresponding comprehensive weights corresponding to an optimal processing person to the optimal processing person according to an embodiment of the present invention is shown;

[0092] Figure 6 A module diagram of a conference record processing device according to an embodiment of the present invention is shown;

[0093] Figure 7 A schematic diagram showing the structure of a computer device suitable for implementing an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0094] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0095] The terms “first,” “second,” etc. used herein do not particularly refer to an order or sequence, nor are they used to limit the present invention. They are only used to distinguish elements or operations described with the same technical terms.

[0096] The words “include,” “including,” “have,” “contain,” etc. used in this document are open-ended terms, meaning including but not limited to.

[0097] As used herein, "and / or" includes any and all combinations of the items mentioned.

[0098] It should be noted that the acquisition, storage, use, and processing of data in the technical solution of the present invention are in compliance with the relevant provisions of national laws and regulations.

[0099] The embodiment of the present invention discloses a method for processing conference records, such as Figure 1 As shown, the method specifically includes the following steps:

[0100] S101: According to a current meeting record, determine a similar record with the highest similarity from multiple historical meeting records; and based on the similar records, determine the optimal person to handle the current meeting record.

[0101] S102: Determine, based on sources of multiple current meeting records, current meeting record sets corresponding to different sources; and obtain a comprehensive weight of all current meeting records based on the weights of the moderators corresponding to the current meeting records in the current meeting record set.

[0102] S103: Determine the pending records corresponding to each optimal processing personnel based on the correspondence between the current meeting record and the optimal processing personnel; send the multiple pending records corresponding to the optimal processing personnel and the corresponding comprehensive weights to the optimal processing personnel, so that the optimal processing personnel processes the corresponding pending records based on the comprehensive weights.

[0103] The meeting record processing method and device provided by the present invention determine the similar record with the highest similarity from multiple historical meeting records based on the current meeting record; and determine the optimal processing personnel of the current meeting record based on the similar records. It can determine the similar record closest to the current meeting record from the actual historical meeting records based on the similarity that fully reflects the degree of similarity between the meeting records, and fully considers the rule that if a processing personnel has processed a historical meeting record, then when processing the current meeting record similar to the historical meeting record, the processing proficiency is higher. Therefore, the above steps can not only improve the accuracy of determining similar records, but also improve the accuracy of determining the optimal processing personnel. Moreover, the above steps are relatively simple, with low computational complexity, and can be automatically executed with low dependence on manual labor. Therefore, the above steps can not only improve the speed of determining similar records, but also improve the speed of determining the optimal processing personnel, so that the above steps can improve the accuracy and speed of the overall meeting record processing. By determining a current meeting record set corresponding to different sources based on the sources of multiple current meeting records; and obtaining a comprehensive weight of all current meeting records based on the host weight corresponding to the current meeting record in the current meeting record set, it can fully consider the impact of the source of the meeting record on the importance of the meeting record, and the rule that the importance of the host corresponding to the meeting record is positively correlated with the importance of the meeting record, and realize the comprehensive weight of the corresponding meeting record based on the source and the host weight that can measure the importance of the host, so that the determined comprehensive weight can accurately reflect the importance of the meeting record, and the importance is positively correlated with the priority, so the determined comprehensive weight can also accurately reflect the priority of the meeting record, so the above steps can improve the accuracy of the determined comprehensive weight, and the above steps can form a set, and can perform related operations and processing with the set as the processing unit, indirectly reducing the amount of data processed. In addition, the above steps can be automatically executed, with low dependence on manual labor, so it can also improve the speed of determining the comprehensive weight, so the above steps can improve the accuracy and speed of the overall meeting record processing.By determining the pending records corresponding to each optimal processing personnel based on the correspondence between the current meeting records and the optimal processing personnel; sending the multiple pending records corresponding to the optimal processing personnel and the corresponding comprehensive weights to the optimal processing personnel, so that the optimal processing personnel can process the corresponding pending records based on the comprehensive weights, the specific pending meeting records assigned to the specific processing personnel can be suitable for the processing of the processing personnel, which is conducive to enabling the processing personnel to skillfully process the assigned meeting records, and enabling the processing personnel to directly determine the order and method of processing the meeting records based on the accurate comprehensive weights, without the need for the processing personnel to conduct additional analysis of the meeting records, so that the accuracy and speed of the meeting record processing can be directly improved, and the step of assigning meeting records can also be automatically executed, reducing the degree of dependence on manual labor, so that the speed of meeting record processing can be further improved. In summary, the meeting record processing method and device provided by the present invention can improve the speed and accuracy of meeting record processing, thereby improving the efficiency of meeting record processing.

[0104] In an optional embodiment, further comprising:

[0105] Before determining the most similar record from multiple historical meeting records based on the current meeting record,

[0106] Performing voice-to-text processing on the preset current conference voice to obtain a current conference record corresponding to the voice source of the current conference voice;

[0107] A projection set is formed based on multiple preset current conference projection images, and text is extracted from one of the current conference projection images in the projection set to obtain the current conference record corresponding to the projection source of the projection set.

[0108] Exemplarily, the speech-to-text processing of the preset current conference voice to obtain the current conference record corresponding to the voice source of the current conference voice can be, but is not limited to, using existing speech-to-text tools, software, applications or algorithms to process all the current conference voices as a whole to obtain the overall conference text content, and using relevant natural language processing algorithms, tools, programs or applications to sentence-process the overall conference text content to obtain multiple current conference records corresponding to the voice source of the current conference voice, or using existing speech-to-text tools, software, applications or algorithms to separately process the voice of each sentence in the current conference voice to obtain one or more current conference records of the corresponding voice source. It should be noted that the specific implementation method of speech-to-text processing of the preset current conference voice to obtain the current conference record corresponding to the voice source of the current conference voice can be determined by those skilled in the art according to actual circumstances. The above description is only an example and does not constitute a limitation.

[0109] Exemplarily, the text extraction from one of the current conference projection images in the projection set to obtain the current conference record corresponding to the projection source of the projection set can be, but is not limited to, randomly selecting a current conference projection image from the projection set as the target current conference projection image, or selecting the current conference projection image of the current conference projection with the latest appearance time from the projection set as the target current conference projection image, and using existing text extraction tools, algorithms, programs or software, etc. and related natural language processing algorithms, tools, programs or applications, etc. to extract and sentence the text in the target current conference projection image to obtain one or more current conference records of the corresponding projection source. It should be noted that the specific implementation method of extracting text from one of the current conference projection images in the projection set to obtain the current conference record corresponding to the projection source of the projection set can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0110] Through the above steps, the conference voice and conference screen projection can be extracted quickly and accurately to obtain the current meeting records of the corresponding voice source and the current meeting records of the corresponding screen projection source, thereby improving the comprehensiveness and accuracy of the collected current meeting records. Based on the formed set, a projection image is selected from all projection images for processing to obtain the meeting record, which can effectively reduce the number of current meeting records used for subsequent processing, which is conducive to improving the subsequent processing speed, and provides a good foundation for the subsequent steps of performing related processing of current meeting records based on different sources, thereby improving the speed and accuracy of the overall meeting record processing.

[0111] In an optional embodiment, forming a projection set based on multiple preset current conference projection images includes:

[0112] Based on the pixel information of all current conference projection images, the pixel similarity between all current conference projection images is obtained;

[0113] Repeating the step of forming a projection set until there is no current conference projection image that is not included in the projection set, the step of forming a projection set includes:

[0114] Select a target image from the current conference projection images that are not included in the projection set;

[0115] One of the projection sets is formed based on the selected images in the current conference projection images that are not included in the projection set, except the target image, whose pixel similarity with the target image is greater than a preset pixel similarity threshold, and the target image.

[0116] Exemplarily, the pixel similarity between all current conference projection images is obtained based on the pixel information of all current conference projection images. It can be, but is not limited to, obtaining the pixel values ​​of all pixels of each current conference projection image based on the pixel information of all current conference projection images. For any two current conference projection images (all the current conference projection images obtained must have the same image size and number of pixels), the pixel values ​​of each corresponding pixel position can be compared to obtain the number of identical pixels between them, and then the number of identical pixels can be divided by the number of pixels of any current conference projection image to obtain the pixel similarity between them. Among them, the pixel similarity obtained is specifically obtained by performing pairwise calculations on all current conference projection images. For example, if there is a current conference projection image A, a current conference projection image B, a current conference projection image C, and a current conference projection image D, then the pixel similarity between all current conference projection images specifically includes the pixel similarity between A and B, the pixel similarity between A and C, the pixel similarity between A and D, the pixel similarity between B and C, the pixel similarity between B and D, and the pixel similarity between C and D. It should be noted that the specific implementation method for obtaining the pixel similarity between all current conference projection images based on the pixel information of all current conference projection images can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0117] Exemplarily, the selection of a target image from the current conference projection images that have not been included in the projection set may be, but is not limited to, selecting the current conference projection image with the earliest corresponding projection appearance time from the current conference projection images that have not been included in the projection set as the target image, or randomly selecting a current conference projection image as the target image. It should be noted that the specific implementation method of selecting a target image from the current conference projection images that have not been included in the projection set can be determined by those skilled in the art based on actual conditions. The above description is only an example and does not constitute a limitation thereto.

[0118] Exemplarily, the preset pixel similarity threshold may be, but is not limited to, 80%. It should be noted that the preset pixel similarity threshold may be determined by those skilled in the art based on actual conditions. The above description is for example only and does not constitute a limitation thereto.

[0119] Exemplarily, the steps of repeatedly performing the steps of forming a projection set are as follows:

[0120] There are current conference projection image A, current conference projection image B, current conference projection image C and current conference projection image D, and the preset pixel similarity threshold is set to 80%.

[0121] Initially, there is no projection set, so A, B, C, and D are all current conference projection images that are not included in the projection set. At this time, A is selected as the target image. Then, among the current conference projection images that are not included in the projection set, the other images except the target image are B, C, and D. The pixel similarity between A and B is 40%, the pixel similarity between A and C is 81%, and the pixel similarity between A and D is 20%. Since the pixel similarity between A and C of 81% is greater than the preset pixel similarity threshold of 80%, the candidate image is determined to be C. At this time, one of the projection sets {A, C} is formed based on the target image A and the candidate image C.

[0122] At this time, there is a projection set {A, C}, as well as the current conference projection images B and D that are not included in the projection set, and B is selected as the target image; therefore, among the current conference projection images that are not included in the projection set, the other images except the target image are D, and the pixel similarity between B and D is 91%. Since the pixel similarity between B and D is 91% greater than the preset pixel similarity threshold of 91%, the image to be selected is determined to be D. At this time, one of the projection sets {B, D} is formed based on the target image B and the image to be selected D; it is worth mentioning that if the pixel similarity between B and D is less than the preset pixel similarity threshold, B and D can be used as the only elements of the set respectively, and two projection sets {B} and {D} are obtained.

[0123] Now, there are projection sets {A, C} and {B, D}, and there are no current conference projection images that are not included in the projection sets, so the iteration ends.

[0124] It should be noted that the specific implementation method of repeatedly executing the steps of forming a projection set can be determined by those skilled in the art based on actual conditions. The above description is only an example and does not constitute a limitation to this.

[0125] Since the distribution of pixels determines the characteristics and appearance of the image, the pixel similarity can fully and accurately reflect the degree of similarity between images. Therefore, the above steps can cluster similar current conference projection images to obtain a projection set based on the pixel similarity that fully and accurately reflects the degree of similarity between images, so that the current conference projection images in the projection set are very similar. Therefore, when selecting projection images from the projection set later, no matter which one is selected, the extracted meeting records are not much different. This is beneficial to effectively reduce the number of current meeting records for subsequent processing, and reduce the degree of omission when extracting meeting records with different features, thereby improving the comprehensiveness, accuracy and speed of related processing based on the extracted current meeting records in subsequent steps, thereby improving the speed and accuracy of the overall meeting record processing.

[0126] In an optional embodiment, forming a projection set based on multiple preset current conference projection images includes:

[0127] Extract text from all current conference projection images to obtain the corresponding text content;

[0128] Based on the text content, obtaining the text similarity between all corresponding current conference projection images;

[0129] Repeating the step of forming a projection set until there is no current conference projection image that is not included in the projection set, the step of forming a projection set includes:

[0130] Select a target image from the current conference projection images that are not included in the projection set;

[0131] One of the projection sets is formed based on the selected images in the current conference projection images that are not included in the projection set, except the target image, whose text similarity with the target image is greater than a preset text similarity threshold, and the target image.

[0132] Exemplarily, the text extraction from all current conference projection images to obtain the corresponding text content may be performed by, but is not limited to, using existing text extraction tools, software, applications, or algorithms to process the current conference projection images to obtain the corresponding text content. It should be noted that the specific implementation method for extracting text from all current conference projection images to obtain the corresponding text content can be determined by those skilled in the art based on actual conditions. The above description is for example only and does not constitute a limitation thereto.

[0133] Exemplarily, the text similarity between all the corresponding current conference projection images is obtained based on the text content. For any two current conference projection images, their text content can be used as input (one projection image corresponds to one text content), and the text similarity between them is determined by an existing text similarity algorithm (for example, but not limited to the TF-IDF algorithm). The text similarity obtained is specifically obtained by calculating the text content corresponding to all the current conference projection images in pairs. For example, if there are current conference projection images A, current conference projection images B, current conference projection images C and current conference projection images D, then the text similarity between all the current conference projection images specifically includes the text similarity between A and B, the text similarity between A and C, the text similarity between A and D, the text similarity between B and C, the text similarity between B and D, and the text similarity between C and D. It should be noted that the specific implementation method for obtaining the text similarity between all the corresponding current conference projection images based on the text content can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation.

[0134] Exemplarily, the selection of a target image from the current conference projection images that have not been included in the projection set can be, but is not limited to, selecting the current conference projection image with the earliest corresponding projection appearance time from the current conference projection images that have not been included in the projection set as the target image, or randomly selecting a current conference projection image as the target image. It should be noted that the specific implementation method of selecting a target image from the current conference projection images that have not been included in the projection set can be determined by those skilled in the art based on actual conditions. The above description is only an example and does not constitute a limitation thereto.

[0135] Exemplarily, the preset text similarity threshold may be, but is not limited to, 80%. It should be noted that the preset text similarity threshold may be determined by those skilled in the art based on actual conditions, and the above description is for example only and does not constitute a limitation thereto.

[0136] Exemplarily, the steps of repeatedly performing the steps of forming a projection set are as follows:

[0137] There are current conference screen projection image A, current conference screen projection image B, current conference screen projection image C and current conference screen projection image D, and the preset text similarity threshold is set to 80%.

[0138] Initially, there is no projection set, so A, B, C, and D are all current conference projection images that are not included in the projection set. At this time, A is selected as the target image. Then, among the current conference projection images that are not included in the projection set, the other images except the target image are B, C, and D. The text similarity between A and B is 40%, the text similarity between A and C is 81%, and the text similarity between A and D is 20%. Since the text similarity between A and C of 81% is greater than the preset text similarity threshold of 80%, the candidate image is determined to be C. At this time, one of the projection sets {A, C} is formed based on the target image A and the candidate image C.

[0139] At this time, there is a projection set {A, C}, as well as current conference projection images B and D that are not included in the projection set, and B is selected as the target image; therefore, among the current conference projection images that are not included in the projection set, the other images except the target image are D, and the text similarity between B and D is 91%. Since the text similarity between B and D of 91% is greater than the preset text similarity threshold of 91%, the image to be selected is determined to be D. At this time, one of the projection sets {B, D} is formed based on the target image B and the image to be selected D; it is worth mentioning that if the text similarity between B and D is less than the preset text similarity threshold, B and D can be used as the only elements of the set respectively, and two projection sets {B} and {D} are obtained.

[0140] Now, there are projection sets {A, C} and {B, D}, and there are no current conference projection images that are not included in the projection sets, so the iteration ends.

[0141] It should be noted that the specific implementation method of repeatedly executing the steps of forming a projection set can be determined by those skilled in the art based on actual conditions. The above description is only an example and does not constitute a limitation to this.

[0142] Since the text content determines the characteristics and appearance of the image, and the meeting records extracted subsequently are also mainly text, the text similarity can fully and accurately reflect the similarity between images, that is, the similarity between the potential meeting records in the projection image. Therefore, the above steps can be based on the text similarity that fully and accurately reflects the similarity between images, that is, the similarity between the potential meeting records in the projection image, and cluster similar current meeting projection images to obtain a projection set, so that the current meeting projection images in the projection set are very similar. Therefore, when selecting a projection image from the projection set subsequently, no matter which one is selected, the extracted meeting records are not much different. This is beneficial to reduce the degree of omission when extracting meeting records with different features on the basis of effectively reducing the number of current meeting records for subsequent processing, thereby improving the comprehensiveness, accuracy and speed of related processing based on the extracted current meeting records in subsequent steps, thereby improving the speed and accuracy of the overall meeting record processing.

[0143] In an optional embodiment, if Figure 2 As shown, the method of determining the similar record with the highest similarity from multiple historical meeting records based on the current meeting record includes the following steps:

[0144] S201: performing word segmentation processing on the current meeting record to obtain corresponding current words, and performing word segmentation processing on a plurality of historical meeting records to obtain historical words corresponding to the historical meeting records.

[0145] S202: Determine the record similarity between the current meeting record and each of the historical meeting records based on the current vocabulary and the historical vocabulary corresponding to each of the historical meeting records.

[0146] S203: Determine the historical conference record with the highest record similarity as the similar record.

[0147] Exemplarily, a current meeting record corresponds to multiple current words.

[0148] Exemplarily, the word segmentation processing of the current meeting record to obtain the corresponding current vocabulary may be, but is not limited to, inputting the current meeting record into an existing word segmentation algorithm, word segmentation tool, word segmentation application or word segmentation model, etc., for processing to obtain the corresponding current vocabulary, wherein the properties of the current vocabulary include but are not limited to verbs, nouns, adjectives, subjects, predicates, objects, attributives, adverbials and complements, etc. It should be noted that the specific implementation method for performing word segmentation processing on the current meeting record to obtain the corresponding current vocabulary and the properties of the current vocabulary can be determined by those skilled in the art according to actual circumstances. The above description is only an example and does not constitute a limitation thereto.

[0149] For example, a historical meeting record corresponds to multiple historical words.

[0150] Exemplarily, the specific implementation method of performing word segmentation processing on the multiple historical meeting records to obtain historical vocabulary corresponding to the historical meeting records can refer to the description of the step of performing word segmentation processing on the current meeting record to obtain the corresponding current vocabulary in the embodiment of the present invention, which will not be repeated here.

[0151] For example, there is a record similarity between a current meeting record and a historical meeting record.

[0152] Exemplarily, the historical meeting record with the highest record similarity is specifically the historical meeting record with the highest record similarity for a current meeting record. For example, there are current meeting record A, historical meeting record B, historical meeting record C, and historical meeting record D. The record similarity between A and B is 50%, the record similarity between A and C is 70%, and the record similarity between A and D is 60%. Then, the record similarity between A and C is the highest, and the historical meeting record with the highest record similarity is C.

[0153] Through steps S201 to S203, the parameter granularity for determining record similarity can be refined to the specific words that constitute the record, thereby improving the accuracy of determining record similarity, thereby improving the accuracy of determining similar records, and further improving the accuracy of subsequent determination of the optimal processing personnel based on similar records, thereby improving the accuracy of the overall meeting record processing.

[0154] In an optional embodiment, determining the record similarity between the current meeting record and each of the historical meeting records based on the current vocabulary and the historical vocabulary corresponding to each of the historical meeting records includes:

[0155] Taking the intersection of the current vocabulary and the historical vocabulary corresponding to the historical meeting records to obtain a vocabulary intersection;

[0156] Taking a union of the current vocabulary and the historical vocabulary corresponding to the historical meeting records to obtain a vocabulary union;

[0157] Based on the number of words in the vocabulary intersection and the number of words in the vocabulary union, the record similarity between the current meeting record and the historical meeting record is obtained.

[0158] Exemplarily, the intersection of the current vocabulary and the historical vocabulary corresponding to the historical meeting records to obtain a vocabulary intersection can be, but is not limited to, clustering multiple current vocabulary corresponding to the current meeting record to obtain a current vocabulary vector or current vocabulary set including all current vocabulary of the current meeting record, and clustering multiple historical vocabulary corresponding to each historical meeting record to obtain a historical vocabulary vector or historical vocabulary set including all historical vocabulary of the historical meeting record (one historical meeting record corresponds to one historical vocabulary vector or one historical vocabulary set), and performing an intersection operation between the current vocabulary vector or current vocabulary set and a corresponding historical vocabulary vector or historical vocabulary set to obtain a corresponding vocabulary intersection (one current meeting record and one historical meeting record correspond to one vocabulary intersection). It should be noted that the specific implementation method of taking the intersection of the current vocabulary and the historical vocabulary corresponding to the historical meeting records to obtain a vocabulary intersection can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0159] Exemplarily, the union of the current vocabulary and the historical vocabulary corresponding to the historical meeting records to obtain a vocabulary union can be, but is not limited to, clustering the multiple current vocabulary corresponding to the current meeting record to obtain a current vocabulary vector or current vocabulary set including all current vocabulary of the current meeting record, and clustering the multiple historical vocabulary corresponding to each historical meeting record to obtain a historical vocabulary vector or historical vocabulary set including all historical vocabulary of the historical meeting record (one historical meeting record corresponds to one historical vocabulary vector or one historical vocabulary set), and performing a union operation between the current vocabulary vector or current vocabulary set and a corresponding historical vocabulary vector or historical vocabulary set to obtain a corresponding vocabulary union (one current meeting record and one historical meeting record correspond to one vocabulary union). It should be noted that the specific implementation method of taking the union of the current vocabulary and the historical vocabulary corresponding to the historical meeting records to obtain a vocabulary union can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0160] For example, since the set elements in the vocabulary intersection and the vocabulary union are all words of the meeting minutes, the number of words in the vocabulary intersection and the number of words in the vocabulary union can be directly known.

[0161] Exemplarily, based on the number of words in the vocabulary intersection and the number of words in the vocabulary union, the record similarity between the current meeting record and the historical meeting record is obtained, which can be achieved by, but not limited to, a cosine similarity algorithm, and is specifically expressed as, but not limited to, the following formula:

[0162]

[0163] Among them, |N(u)∩N(v)| represents the number of words in the vocabulary intersection between the current meeting record u and the historical meeting record v, |N(u)||N(v)| represents the number of words in the vocabulary union between the current meeting record u and the historical meeting record v, and N(u) represents all current words corresponding to the current meeting record u, N(v) represents all historical words corresponding to the historical meeting record v, and w uv Indicates the similarity between the current meeting record u and the historical meeting record v.

[0164] It should be noted that the specific implementation method for obtaining the record similarity between the current meeting record and the historical meeting record based on the number of words in the vocabulary intersection and the number of words in the vocabulary union can be determined by technical personnel in this field according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0165] Since the intersection and union between two objects are the mainstream input parameters for determining similarity between them, the above steps determine the intersection and union, and use the intersection and union as input to calculate the corresponding record similarity. This can improve the accuracy of the obtained record similarity, thereby improving the accuracy of subsequent determination of similar records, and thus improving the accuracy of overall meeting record processing. Furthermore, the computational complexity of the process of determining the intersection and union, and determining similarity based on the intersection and union, is low, thus speeding up the determination of record similarity and, consequently, the overall speed of meeting record processing.

[0166] In an optional embodiment, if Figure 3 As shown, determining the optimal person to handle the current meeting record based on the similar records includes the following steps:

[0167] S301: Based on the similar records, corresponding historical processing information is obtained.

[0168] S302: Determine multiple processing times corresponding to each historical processing person according to the historical processing information.

[0169] S303: Based on the current time and the processing time, obtain the time differences corresponding to multiple processing times of each historical processing personnel, and determine the maximum time difference based on all the time differences corresponding to all historical processing personnel.

[0170] S304: Obtaining a processing weight corresponding to the historical processing personnel according to the multiple time differences corresponding to the historical processing personnel, the maximum time difference, and a preset correction coefficient.

[0171] S305: Determine the historical processing personnel with the largest processing weight as the optimal processing personnel.

[0172] Exemplarily, the historical processing information includes but is not limited to multiple processing records for the similar record (the nature of the similar record is a historical meeting record) within a preset time period, and each processing record at least includes the corresponding historical processing personnel for that processing (if multiple historical processing personnel participated in that processing, one of the historical processing personnel is selected as the historical processing personnel for that processing) and the processing time. Among them, the preset time period may be but is not limited to the past 365 days or the past 1095 days, etc. The preset time period may be determined by those skilled in the art according to actual conditions, and the embodiments of the present invention do not impose any restrictions on this. It should be noted that the specific content of the historical processing information may be determined by those skilled in the art according to actual conditions, and the above description is only an example and does not constitute a limitation on this.

[0173] Exemplarily, obtaining the corresponding historical processing information based on the similar records may include, but is not limited to, using the similar records as input and querying a related system, database, or application to obtain the historical processing information corresponding to the similar records. It should be noted that the specific implementation of obtaining the corresponding historical processing information based on the similar records can be determined by those skilled in the art based on actual circumstances, and the above description is merely an example and does not constitute a limitation thereto.

[0174] Exemplarily, based on the historical processing information, multiple processing times corresponding to each historical processing personnel are determined. This can be, but is not limited to, parsing the historical processing information to obtain multiple processing records, and determining the corresponding historical processing personnel and processing time based on the processing records. Then, based on the historical processing personnel and processing time corresponding to all processing records, multiple processing times corresponding to each historical processing personnel are determined (because a historical meeting record may be processed multiple times by one person at different times, so a historical processing personnel corresponding to a historical meeting record will have multiple processing times). It should be noted that the specific implementation method for determining multiple processing times corresponding to each historical processing personnel based on the historical processing information can be determined by those skilled in the art according to actual circumstances. The above description is only an example and does not constitute a limitation thereto.

[0175] Exemplarily, the time difference corresponding to multiple processing times of each historical processing personnel is obtained based on the current time and the processing time. It can be, but is not limited to, taking the time difference between the current time and a certain processing time as the time difference corresponding to the processing time, that is, the time difference corresponding to multiple processing times of each historical processing personnel can be obtained (equivalent to one historical meeting record corresponding to multiple processing times of multiple processing records, and each processing time corresponds to a time difference).

[0176] For example, the maximum time difference determined based on all time differences corresponding to all historical processing personnel may be, but is not limited to, selecting a maximum time difference from the time differences corresponding to all processing times corresponding to the corresponding historical meeting records (including the processing times corresponding to each historical processing personnel) to determine as the maximum time difference. Each similar record (being a historical meeting record) corresponds to one maximum time difference.

[0177] Exemplarily, the processing weight corresponding to the historical processing personnel is obtained according to the multiple time differences corresponding to the historical processing personnel, the maximum time difference, and the preset correction coefficient, which can be expressed as but not limited to the following formula:

[0178]

[0179] Among them, p uj represents the processing weight of the historical processing personnel for the similar record, k represents the number of time differences corresponding to the historical processing personnel (for the similar record, one historical processing personnel corresponds to multiple processing times, and one processing time corresponds to one time difference, so one historical processing personnel corresponds to multiple time differences), h represents the maximum time difference, l represents the correction coefficient, d i The specific value of the correction coefficient can be determined by those skilled in the art according to actual conditions, and the embodiment of the present invention does not limit this. For example, the correction coefficient can be, but is not limited to, a coefficient greater than or equal to 1.

[0180] Since historical processing information can reflect the historical processing of similar records, and the relevant time difference can accurately reflect the number of times the corresponding processing personnel processed similar records and the degree of forgetfulness in processing the similar records, the larger the overall time difference, the more the processing personnel has forgotten the method of processing the similar records, and thus the less skilled the processing personnel is, which in turn means that the processing personnel will be less skilled in processing the current meeting records. The more time differences a processing personnel has, the more skilled the processing personnel is in processing the similar records, which in turn means that the processing personnel will be more skilled in processing the current meeting records. Therefore, the number of processing times and the degree of forgetfulness are closely related to the processing personnel's proficiency in processing the current meeting records, that is, they are closely related to the accuracy of processing the current meeting records. Therefore, the above steps can determine the processing weight corresponding to the historical processing personnel based on the historical processing situation and time difference, so that the processing weight can accurately measure the processing quality and processing level of the processing personnel for the current meeting records, thereby measuring the processing accuracy of the processing personnel for the current meeting records, and thus can achieve the determination of the historical processing personnel with the highest processing accuracy as the optimal processing personnel, thereby improving the accuracy of determining the optimal processing personnel, thereby improving the accuracy of meeting record processing.

[0181] In an optional embodiment, determining, based on the sources of the plurality of current meeting records, current meeting record sets corresponding to different sources respectively includes:

[0182] The current conference record whose source is the screen projection source is determined as the current conference screen projection record, and the current conference record whose source is the voice source is determined as the current conference voice record;

[0183] Determining a first current conference record set corresponding to a screen projection source according to the current conference screen projection record;

[0184] A second current conference record set corresponding to the voice source is determined based on the current conference voice record.

[0185] Through the above steps, the meeting record set corresponding to the screen projection source and the meeting record set corresponding to the voice source can be determined more finely and specifically, which is conducive to making the nature of the divided meeting record set clearer, and further conducive to improving the accuracy and speed of the steps related to meeting record processing based on the meeting record set.

[0186] In an optional embodiment, determining the first current conference record set corresponding to the screen projection source based on the current conference screen projection record includes:

[0187] Perform word segmentation on each of the current conference screen projection records to obtain corresponding screen projection vocabulary;

[0188] Determining a first current similarity between all current conference screen projection records based on the screen projection vocabulary corresponding to the current conference screen projection record;

[0189] Repeating the step of forming a first current conference record set until there is no current conference screen projection record that is not included in the first current conference record set, the step of forming the first current conference record set includes:

[0190] Selecting a first target record from the current conference projection records that are not included in the first current conference record set;

[0191] Based on the first candidate record and the first target record in the current conference projection records that are not included in the first current conference record set, the first current similarity between the other records except the first target record and the first target record is greater than the preset current similarity threshold, one of the first current conference record sets is formed.

[0192] Exemplarily, the specific implementation method of segmenting each of the current meeting projection records to obtain the corresponding projection vocabulary can refer to the description of the steps of segmenting the current meeting records to obtain the corresponding current vocabulary in the embodiment of the present invention, which will not be repeated here.

[0193] Exemplarily, a current conference screen projection record corresponds to multiple screen projection words.

[0194] Exemplarily, the first current similarity between all current conference screen projection records is determined based on the screen projection vocabulary corresponding to the current conference screen projection record. It can be, but is not limited to, for any two current conference screen projection records, their respective corresponding screen projection vocabulary can be used as input to determine the first current similarity between them. The first current similarity obtained is specifically obtained by calculating all current conference screen projection records in pairs. For example, if there are current conference screen projection records A, current conference screen projection records B, current conference screen projection records C and current conference screen projection records D, then the first current similarity between all current conference screen projection records specifically includes the first current similarity between A and B, the first current similarity between A and C, the first current similarity between A and D, the first current similarity between B and C, the first current similarity between B and D, and the first current similarity between C and D. It should be noted that the specific implementation method for determining the first current similarity between all current conference screen projection records based on the screen projection vocabulary corresponding to the current conference screen projection record can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0195] Exemplarily, the selecting a first target record from the current conference screen projection records that have not been included in the first current conference record set may be, but is not limited to, selecting the current conference screen projection record with the earliest corresponding projection appearance time from the current conference screen projection records that have not been included in the first current conference record set as the first target record, or randomly selecting a current conference screen projection record as the first target record. It should be noted that the specific implementation method of selecting a first target record from the current conference screen projection records that have not been included in the first current conference record set can be determined by those skilled in the art based on actual conditions. The above description is only an example and does not constitute a limitation to this.

[0196] Exemplarily, the preset current similarity threshold may be, but is not limited to, a value in the range of (0, 1), such as 80%, 70% or 85%. It should be noted that the preset current similarity threshold may be determined by those skilled in the art based on actual conditions, and the above description is for example only and does not constitute a limitation thereto.

[0197] Exemplarily, the step of repeatedly performing the step of forming the first current conference record set includes the following examples:

[0198] There are current conference screen projection record A, current conference screen projection record B, current conference screen projection record C and current conference screen projection record D, and the preset current similarity threshold is set to 80%.

[0199] Initially, there is no first current conference record set, so A, B, C, and D are all current conference projection records that are not included in the first current conference set. At this time, A is selected as the first target record. Then, among the current conference projection records that are not included in the projection set, the other records except the first target record are B, C, and D. The first current similarity between A and B is 40%, the first current similarity between A and C is 81%, and the first current similarity between A and D is 20%. Since the first current similarity of 81% between A and C is greater than the preset current similarity threshold of 80%, the first candidate record is determined to be C. At this time, one of the first current conference record sets {A, C} is formed based on the first target record A and the first candidate record C.

[0200] At this time, there is a first current meeting record set {A, C}, as well as current meeting screen projection records B and D that are not included in the first current meeting record set, and B is selected as the first target record; therefore, the other records in the current meeting screen projection records that are not included in the first current meeting record set except the first target record are D, and the first current similarity between B and D is 91%. Since the first current similarity of 91% between B and D is greater than the preset current similarity threshold of 91%, the first candidate record is determined to be D. At this time, one of the first current meeting record sets {B, D} is formed based on the first target record B and the first candidate record D; it is worth mentioning that if the first current similarity between B and D is less than the preset current similarity threshold, B and D can be used as the only elements of the set respectively, and two first current meeting record sets {B} and {D} are obtained.

[0201] Now, there exists a first current conference record set {A, C} and {B, D}, and there is no current conference screen projection record that is not included in the first current conference record set, so the iteration ends.

[0202] It should be noted that the specific implementation method of repeatedly executing the step of forming the first current meeting record set can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation thereto.

[0203] Since the vocabulary corresponding to the relevant meeting records can accurately and intuitively reflect the nature and characteristics of the meeting records, the above steps will refine the granularity of the basis for determining the first current meeting record set to the vocabulary, and determine the first current similarity between all current meeting screen projection records based on the vocabulary, which can make the first current similarity more accurately reflect the degree of similarity between the current meeting screen projection records, thereby improving the accuracy of the subsequent formation of multiple first current meeting record sets based on the similarity. By repeatedly executing the steps of forming the first current meeting record set, multiple current meeting screen projection records with a high degree of similarity can be clustered separately, so that the multiple current meeting screen projection records in each first current meeting record set are relatively similar. Therefore, it can improve the accuracy of subsequent determination of relevant weights based on the set, and is beneficial to ensuring that when multiple records in a set are given the same weight in subsequent steps, they have higher accuracy, and can ensure that when representative records are selected from the first current meeting record set subsequently, no matter which one is selected, there is little difference, which is beneficial to reducing the degree of omission when extracting records to be processed with different characteristics on the basis of effectively reducing the number of records to be processed for subsequent processing, thereby helping to improve the comprehensiveness, accuracy and speed of related processing based on the extracted records to be processed in subsequent steps, thereby improving the speed and accuracy of the overall meeting record processing.

[0204] In an optional embodiment, determining the first current similarity between all current conference screen projection records based on the screen projection vocabulary corresponding to the current conference screen projection record includes:

[0205] Intersections of two screen projection words corresponding to all the current conference screen projection records are taken to obtain a plurality of corresponding first intersections;

[0206] Taking the union of two screen projection words corresponding to all the current conference screen projection records to obtain multiple corresponding first unions;

[0207] Based on the number of words in the first intersection and the number of words in the first union, a first current similarity between the corresponding two current conference projection records is obtained.

[0208] Exemplarily, the intersection of the screen projection words corresponding to all the current conference screen projection records is taken in pairs to obtain multiple corresponding first intersections. This can be, but is not limited to, clustering the multiple screen projection words corresponding to the current conference screen projection record to obtain a screen projection word vector or a screen projection word set including all the screen projection words of the current conference screen projection record (one current conference screen projection record corresponds to one screen projection word vector or screen projection word set), and then the screen projection word vectors or screen projection word sets corresponding to different current conference screen projection records are taken in pairs to obtain multiple corresponding first intersections (one current conference screen projection record and another current conference screen projection record correspond to one first intersection). It should be noted that the specific implementation method of taking the intersection of the screen projection words corresponding to all the current conference screen projection records in pairs to obtain multiple corresponding first intersections can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0209] Exemplarily, the screen projection words corresponding to all the current conference screen projection records are taken as unions in pairs to obtain multiple corresponding first unions. This can be, but is not limited to, clustering the multiple screen projection words corresponding to the current conference screen projection record to obtain a screen projection word vector or a screen projection word set including all the screen projection words of the current conference screen projection record (one current conference screen projection record corresponds to one screen projection word vector or screen projection word set), and then the screen projection word vectors or screen projection word sets corresponding to different current conference screen projection records are taken as unions in pairs to obtain multiple corresponding first unions (one current conference screen projection record and another current conference screen projection record correspond to one first union). It should be noted that the specific implementation method of taking the screen projection words corresponding to all the current conference screen projection records as unions in pairs to obtain multiple corresponding first unions can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0210] For example, since the set elements in the first intersection and the first union are all screen projection words of the current conference screen projection record, the number of words in the first intersection and the number of words in the first union can be directly known.

[0211] Exemplarily, based on the number of words in the first intersection and the number of words in the first union, the first current similarity between the corresponding two current conference projection records is obtained, which can be achieved by, but not limited to, a cosine similarity algorithm. As for the specific principles of the cosine similarity algorithm, reference can be made to the description of the steps of obtaining the record similarity between the current meeting record and the historical meeting record based on the number of words in the vocabulary intersection and the number of words in the vocabulary union in the embodiment of the present invention, which will not be repeated here. It should be noted that the specific implementation method of obtaining the first current similarity between the corresponding two current conference projection records based on the number of words in the first intersection and the number of words in the first union can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0212] Since the intersection and union between two objects are the mainstream input parameters for determining the similarity between them, the above steps determine the intersection and union, and use the intersection and union as input to calculate the corresponding first current similarity. This can improve the accuracy of the obtained first current similarity, thereby improving the accuracy of the subsequent determination of the first current meeting record set, and further improving the accuracy of the overall meeting record processing. Furthermore, the computational complexity of the calculation process for determining the intersection and union, and determining the similarity based on the intersection and union, is relatively low, thus speeding up the determination of the first current similarity, thereby speeding up the overall meeting record processing.

[0213] In an optional embodiment, determining a second current conference record set corresponding to a voice source based on the current conference voice record includes:

[0214] Perform word segmentation processing on each of the current conference voice records to obtain corresponding voice vocabulary;

[0215] determining a second current similarity between all current conference voice records based on the voice vocabulary corresponding to the current conference voice record;

[0216] Repeating the step of forming a second current conference record set until there is no current conference voice record that is not included in the second current conference record set, the step of forming the second current conference record set includes:

[0217] Selecting a second target record from the current conference voice records that are not included in the second current conference record set;

[0218] One of the second current meeting record sets is formed based on the second candidate records and the second target record, whose second current similarity with the second target record in other records of the current meeting voice records other than the second target record is greater than the preset current similarity threshold.

[0219] Exemplarily, the specific implementation method of segmenting each of the current conference voice records to obtain corresponding voice vocabulary can refer to the description of the steps of segmenting each of the current conference screen projection records to obtain corresponding screen projection vocabulary in the embodiment of the present invention, which will not be repeated here.

[0220] Exemplarily, the specific implementation method of determining the second current similarity between all current conference voice records based on the voice vocabulary corresponding to the current conference voice record can refer to the description of the step of determining the first current similarity between all current conference screen projection records based on the screen projection vocabulary corresponding to the current conference screen projection record in the embodiment of the present invention, which will not be repeated here.

[0221] Illustratively, a current conference voice record corresponds to multiple voice words.

[0222] Exemplarily, the specific implementation method of selecting a second target record from the current conference voice record that has not been included in the second current conference record set can refer to the description of the steps of selecting a first target record from the current conference screen projection record that has not been included in the first current conference record set in the embodiment of the present invention, which will not be repeated here.

[0223] For example, the specific implementation of the step of repeatedly executing to form the second current conference record set can refer to the description of the step of repeatedly executing to form the first current conference record set in the embodiment of the present invention, which will not be repeated here.

[0224] Since the vocabulary corresponding to the relevant meeting records can accurately and intuitively reflect the nature and characteristics of the meeting records, the above steps will refine the granularity of the basis for determining the second current meeting record set to the vocabulary, and determine the second current similarity between all current meeting voice records based on the vocabulary, which can make the second current similarity more accurately reflect the degree of similarity between the current meeting voice records, thereby improving the accuracy of the subsequent formation of multiple second current meeting record sets based on the similarity. By repeatedly executing the steps of forming the second current meeting record set, multiple current meeting voice records with a high degree of similarity can be clustered separately, so that the multiple current meeting voice records in each second current meeting record set are relatively similar. Therefore, it can improve the accuracy of subsequent determination of relevant weights based on the set, and is beneficial to making multiple records in a set in subsequent steps be given the same weight with higher accuracy, and can also make the subsequent selection of representative records from the second current meeting record set, no matter which one is selected, there will be little difference, which is beneficial to effectively reduce the number of pending records for subsequent processing, and reduce the degree of omission when extracting pending records with different characteristics, thereby helping to improve the comprehensiveness, accuracy and speed of related processing based on the extracted pending records in subsequent steps, thereby improving the speed and accuracy of the overall meeting record processing.

[0225] In an optional implementation, determining the second current similarity between all current conference voice records based on the voice vocabulary corresponding to the current conference voice record includes:

[0226] Intersections of two voice words corresponding to all the current conference voice records are taken to obtain a plurality of corresponding second intersections;

[0227] Taking unions of two voice words corresponding to all the voice records of the current conference to obtain a plurality of corresponding second unions;

[0228] Based on the number of words in the second intersection and the number of words in the second union, a second current similarity between the corresponding two current conference voice records is obtained.

[0229] For example, the specific implementation method of taking the intersection of all the voice words corresponding to the current conference voice records in pairs to obtain multiple corresponding second intersections can refer to the description of the steps of taking the intersection of all the screen projection words corresponding to the current conference screen projection records in pairs to obtain multiple corresponding first intersections in the embodiment of the present invention, which will not be repeated here.

[0230] For example, the specific implementation method of taking the union of all the voice words corresponding to the current conference voice records in pairs to obtain multiple corresponding second unions can refer to the description of the steps of taking the union of all the screen projection words corresponding to the current conference screen projection records in pairs to obtain multiple corresponding first unions in the embodiment of the present invention, which will not be repeated here.

[0231] Exemplarily, since the set elements in the second intersection and the second union are all voice words of the current conference voice record, the number of words in the second intersection and the number of words in the second union can be directly known.

[0232] Exemplarily, the specific implementation method of obtaining the second current similarity between the corresponding two current conference voice records based on the number of words in the second intersection and the number of words in the second union can refer to the description of the steps of obtaining the first current similarity between the corresponding two current conference projection records based on the number of words in the first intersection and the number of words in the first union in the embodiment of the present invention, which will not be repeated here.

[0233] Since the intersection and union between two objects are the mainstream input parameters for determining the similarity between them, the above steps determine the intersection and union, and use the intersection and union as input to calculate the corresponding second current similarity. This can improve the accuracy of the obtained second current similarity, thereby improving the accuracy of the subsequent determination of the second current meeting record set, and further improving the accuracy of the overall meeting record processing. Furthermore, the computational complexity of the calculation process for determining the intersection and union, and determining the similarity based on the intersection and union, is relatively low, thus increasing the speed of determining the second current similarity, thereby increasing the speed of overall meeting record processing.

[0234] In an optional implementation, obtaining the comprehensive weight of all current meeting records based on the host weight corresponding to the current meeting record in the current meeting record set includes:

[0235] Determine whether the source corresponding to the current conference record set is a screen projection source or a voice source,

[0236] If it is a screen projection source, based on the current meeting record, obtain the number of projections of the current meeting projection image included in the corresponding projection set; based on the number of projections corresponding to the current meeting record and the host weight, obtain the corresponding sub-weight; superimpose the sub-weights corresponding to all current meeting records in the current meeting record set to obtain the projection sub-weight corresponding to the current meeting record set, and use the projection sub-weight as the projection sub-weight of each current meeting record in the current meeting record set;

[0237] If the source is voice, the host weights corresponding to all current meeting records in the current meeting record set are superimposed to obtain the voice sub-weight corresponding to the current meeting record set, and the voice sub-weight is used as the voice sub-weight of each current meeting record in the current meeting record set;

[0238] Based on the screen projection weights or voice weights corresponding to all current meeting records, the corresponding comprehensive weight is determined.

[0239] For example, since the nature and corresponding source of the current conference record set have been clarified in the previous steps, it is possible to directly determine whether the source corresponding to the current conference record set is a screen projection source or a voice source.

[0240] For example, one current meeting record set corresponds to multiple current meeting records, and one current meeting record corresponds to one projection set (since the correspondence between meeting records and projection images is known, and the projection set to which the projection image belongs is known, the correspondence between the current meeting record and the projection set is also known). Therefore, the number of projections of the current meeting projection image included in the corresponding projection set can be directly obtained based on the current meeting record.

[0241] For example, one current meeting record corresponds to one projection number.

[0242] Exemplarily, the corresponding sub-weights are obtained based on the number of screen projections and the host weight corresponding to the current meeting record. It can be, but is not limited to, multiplying the number of screen projections by the host weight to obtain the corresponding sub-weights, wherein one current meeting record corresponds to one sub-weight and one host weight. Wherein, when the source of the current meeting record is the screen projection source, the host weight can be, but is not limited to, the screen projector of the screen projection image corresponding to the current meeting record (which can be determined based on the screen projection account) or the corresponding speaker's position weight (the higher the position, the higher the position weight, and the correspondence between the position and the position weight can be preset according to the actual situation), wherein the value of the host weight must be greater than or equal to 1. It should be noted that the specific implementation method of obtaining the corresponding sub-weights based on the number of screen projections and the host weight corresponding to the current meeting record, as well as the nature of the host weight, can be determined by those skilled in the art according to the actual situation. The above description is only an example and does not constitute a limitation to this.

[0243] Exemplarily, when the source of the current meeting record is a voice source, the host weight corresponding to the current meeting record is specifically the weight of the speaker corresponding to the voice information corresponding to the current meeting record, wherein one current meeting record corresponds to one host weight, and the speaker's weight can be, but not limited to, the speaker's position weight (the higher the position, the higher the position weight, and the correspondence between the position and the position weight can be preset according to the actual situation), wherein the value of the host weight must be greater than or equal to 1. The speaker corresponding to the voice information corresponding to the current meeting record can be determined by, but not limited to, a preset voiceprint library. If the voice information cannot find the corresponding speaker in the voiceprint library, the corresponding host weight is defaulted to a preset fixed value (for example, it can be, but not limited to, 1). It should be noted that the source and specific nature of the host weight can be determined by those skilled in the art according to the actual situation. The above description is only an example and does not constitute a limitation to this.

[0244] Through the above steps, the impact of the source nature of the current meeting record on the importance of the current meeting record can be considered in more detail, and the corresponding weights for the current meeting record of the voice source and the current meeting record of the screen projection source are determined based on different inputs, and the overall record characteristics of the set corresponding to the current meeting record are further considered, so that the determined weight of the current meeting record can accurately match the overall importance of multiple similar meeting records, thereby improving the accuracy of the determined weight, thereby improving the accuracy of the determined comprehensive weight, and then improving the accuracy of the overall meeting record processing.

[0245] In an optional embodiment, determining the corresponding comprehensive weight based on the screen projection weights or voice weights corresponding to all current conference records includes:

[0246] Determine whether the current conference record is only a screen projection record of the current conference, only a voice record of the current conference, or both a screen projection record and a voice record of the current conference;

[0247] If only the current conference screen projection record is used, the corresponding comprehensive weight is determined based on the screen projection weight and the preset screen projection coefficient;

[0248] If only the current conference voice record is used, the corresponding comprehensive weight is determined based on the voice sub-weight and the preset voice coefficient;

[0249] If it is both a screen projection record and a voice record of the current meeting, the corresponding first comprehensive weight is determined based on the screen projection weight and the preset screen projection coefficient, and the corresponding second comprehensive weight is determined based on the voice weight and the preset voice coefficient. The first comprehensive weight and the second comprehensive weight are added together to obtain the comprehensive weight.

[0250] For example, since there are inevitably overlapping current meeting records among multiple current meeting records, and among the overlapping current meeting records, some records may be recorded from voice and some from screen projection, then it can be determined that the nature of these overlapping current meeting records is both the current meeting screen projection record and the current meeting voice record. Specifically, the following examples are provided to determine the nature of the current meeting record:

[0251] There are current meeting records A (screen projection), current meeting record B (screen projection), current meeting record C (screen projection), current meeting record D (voice), current meeting record E (voice) and current meeting record F (voice), and the contents of C and D are exactly the same, then the nature of A and B is determined to be only the current meeting screen projection records, the nature of C and D is determined to be both the current meeting screen projection records and the current meeting voice records, and the nature of E and F is determined to be only the current meeting voice records.

[0252] It should be noted that the specific implementation method for determining whether the current meeting record is only the current meeting screen projection record, only the current meeting voice record, or both the current meeting screen projection record and the current meeting voice record can be determined by technical personnel in this field according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0253] Exemplarily, the corresponding comprehensive weight determined based on the projection weight and the preset projection coefficient may be, but is not limited to, multiplying the projection weight by the projection coefficient to obtain the corresponding comprehensive weight, wherein one current meeting record corresponds to one projection weight and one comprehensive weight. The specific value of the projection coefficient can be determined by those skilled in the art according to actual conditions, and the embodiment of the present invention does not impose any restrictions on this. For example, the projection coefficient may be, but is not limited to, a projection coefficient in the range of (0, 1), and may specifically be, but is not limited to, 0.5, 0.6, or 0.55, etc.

[0254] Exemplarily, the corresponding comprehensive weight determined based on the voice sub-weight and the preset voice coefficient may be, but is not limited to, multiplying the voice sub-weight by the voice coefficient to obtain the corresponding comprehensive weight, wherein one current meeting record corresponds to one voice sub-weight and one comprehensive weight. The specific value of the voice coefficient may be determined by those skilled in the art according to actual conditions, and the embodiment of the present invention does not impose any restrictions on this. For example, the voice coefficient may be, but is not limited to, a voice coefficient in the range of (0, 1), specifically, may be, but is not limited to, 0.5, 0.4, or 0.45, etc. Preferably, the sum of the voice coefficient and the projection coefficient is equal to 1.

[0255] Exemplarily, the specific implementation method of determining the corresponding first comprehensive weight based on the screen projection weight and the preset screen projection coefficient can refer to the description of the steps of determining the corresponding comprehensive weight based on the screen projection weight and the preset screen projection coefficient in the embodiment of the present invention, which will not be repeated here.

[0256] Exemplarily, the specific implementation method of determining the corresponding second comprehensive weight based on the voice weight and the preset voice coefficient can refer to the description of the steps of determining the corresponding comprehensive weight based on the voice weight and the preset voice coefficient in the embodiment of the present invention, which will not be repeated here.

[0257] Exemplarily, if it is both a screen projection record and a voice record of the current meeting, the corresponding first comprehensive weight is determined based on the screen projection weight and the preset screen projection coefficient, and the corresponding second comprehensive weight is determined based on the voice weight and the preset voice coefficient. The first comprehensive weight and the second comprehensive weight are added to obtain the comprehensive weight, which can be specifically reflected in but not limited to the following formula:

[0258] f=ar+bs

[0259] Among them, f represents the comprehensive weight for the current meeting record, a represents the voice coefficient, r represents the voice sub-weight, b represents the screen projection coefficient, s represents the screen projection sub-weight, ar represents the second comprehensive sub-weight, and bs represents the first comprehensive sub-weight.

[0260] It should be noted that, for determining the corresponding first comprehensive score weight based on the screen projection score weight and the preset screen projection coefficient, and determining the corresponding second comprehensive score weight based on the voice score weight and the preset voice coefficient, the specific implementation method of adding the first comprehensive score weight and the second comprehensive score weight to obtain the comprehensive weight can be determined by technical personnel in this field according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0261] Through the above steps, we can further consider the impact of different sources of meeting minutes on the importance of meeting minutes, use the screen projection coefficient and voice coefficient to correct the comprehensive weight of meeting minutes from different sources, and further fully consider the rule that if a current meeting minute has been reflected in both the voice source and the screen projection source, it means that its importance is higher. A special comprehensive weight is determined for this current meeting minute. Therefore, the above steps can further refine the parameter granularity for determining the comprehensive weight and further significantly improve the accuracy of the determined comprehensive weight, thereby improving the accuracy of meeting minute processing. Moreover, the relevant calculations and operations of the above steps are relatively simple and have low computational complexity. Therefore, the above steps can also increase the speed of determining the comprehensive weight, thereby improving the speed of overall meeting minute processing.

[0262] In an optional embodiment, if Figure 4 As shown, the process of determining the to-be-processed record corresponding to each optimal processing personnel based on the correspondence between the current meeting record and the optimal processing personnel includes the following steps:

[0263] S401: Based on the correspondence between each current meeting record and the optimal processing personnel in the current meeting record set, determine the number of current meeting records in the current meeting record set corresponding to different optimal processing personnel, and take the optimal processing personnel corresponding to the maximum number of current meeting records as the optimal processing personnel corresponding to the current meeting record set.

[0264] S402: Select a representative record from each current meeting record set, and determine the correspondence between the optimal processing person and the representative record based on the correspondence between the representative record and the current meeting record set and the correspondence between the current meeting record set and the optimal processing person.

[0265] S403: According to the correspondence between the optimal processing personnel and the representative records, all the representative records corresponding to the optimal processing personnel are determined as the records to be processed corresponding to the optimal processing personnel.

[0266] Exemplarily, based on the correspondence between each current meeting record in the current meeting record set and the optimal processing personnel, the number of current meeting records in the current meeting record set corresponding to different optimal processing personnel is determined, and the optimal processing personnel corresponding to the largest number of current meeting records is used as the optimal processing personnel corresponding to the current meeting record set. The following examples are provided:

[0267] There exists a current meeting record set, including current meeting record A, current meeting record B, current meeting record C, current meeting record D, current meeting record E, and current meeting record F. Current meeting record A corresponds to the optimal processing personnel A, current meeting record B corresponds to the optimal processing personnel B, current meeting record C corresponds to the optimal processing personnel B, current meeting record D corresponds to the optimal processing personnel C, current meeting record E corresponds to the optimal processing personnel C, and current meeting record F corresponds to the optimal processing personnel C. It can be seen that A corresponds to 1 (A) of the current meeting record set, B corresponds to 2 (B, C) of the current meeting record set, and C corresponds to 3 (D, E, F) of the current meeting record set. Therefore, the optimal processing personnel corresponding to the maximum current meeting record number of 3 is C, and C is taken as the optimal processing personnel corresponding to the current meeting record set.

[0268] Exemplarily, the selection of a representative record from each current meeting record set may be, but is not limited to, randomly selecting a representative record from each current meeting record set or selecting the current meeting record with the earliest corresponding appearance time (which can be determined by its corresponding voice recording time or corresponding screen projection time, etc.) from each current meeting record set as the representative record. Among them, one current meeting record set corresponds to one representative record. It should be noted that the specific implementation method of selecting a representative record from each current meeting record set can be determined by those skilled in the art according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0269] Exemplarily, a representative record is selected from each current meeting record set, and the correspondence between the representative record and the current meeting record set and the correspondence between the current meeting record set and the optimal processing personnel is determined. Examples include the following:

[0270] The following corresponding relationships exist:

[0271] Representative record A - the first current meeting record set - the best processing person A

[0272] Representative record B - the second current meeting record set - the best processing personnel B

[0273] Representative record C - the third current meeting record set - the best processing personnel C

[0274] Representative record D - the fourth current meeting record set - the best processing personnel A

[0275] Representative record E - the fifth current meeting record set - the best processing person B

[0276] It can be determined that the correspondence between the optimal processing personnel and the representative records is specifically as follows:

[0277] Optimal Processing Personnel A - Representative Record A, Representative Record D

[0278] Optimal Processing Person B - Representative Record B, Representative Record E

[0279] Best Processing Person C - Representative Record C

[0280] It should be noted that for selecting a representative record from each current meeting record set, the specific implementation method of determining the correspondence between the optimal processing personnel and the representative record based on the correspondence between the representative record and the current meeting record set and the correspondence between the current meeting record set and the optimal processing personnel can be determined by technical personnel in this field according to actual conditions. The above description is only an example and does not constitute a limitation to this.

[0281] Through the above steps, the degree of analysis of the correspondence between the current meeting records and the optimal processing personnel can be strengthened, and the records to be processed corresponding to each optimal processing personnel can be determined based on the correspondence with smaller granularity after analysis. This can improve the accuracy of determining the records to be processed corresponding to each optimal processing personnel, thereby improving the accuracy of the overall meeting record processing.

[0282] In an optional embodiment, if Figure 5 As shown, the sending of the plurality of pending records corresponding to the optimal processing personnel and the corresponding comprehensive weights to the optimal processing personnel so that the optimal processing personnel processes the corresponding pending records based on the comprehensive weights includes the following steps:

[0283] S501: sorting the multiple pending records and the corresponding comprehensive weights corresponding to the optimal processing personnel according to the comprehensive weights, and sending them to the optimal processing personnel, so that the optimal processing personnel processes the corresponding pending records based on the sorted comprehensive weights.

[0284] Exemplarily, the sorting by the comprehensive weight may be, but is not limited to, sorting by comprehensive weight from large to small or sorting by comprehensive weight from small to large, etc.

[0285] Through the above steps, the records to be processed can be displayed to the corresponding optimal processing personnel in a clear priority order, so that the optimal processing personnel can more conveniently and intuitively know the best processing order of the records to be processed without having to sort and confirm them personally, reducing the difficulty of the optimal processing personnel in processing meeting records, reducing the time for processing meeting records, and improving the speed and efficiency of processing meeting records.

[0286] Based on the same principle, the embodiment of the present invention discloses a conference record processing device 600, such as Figure 6 As shown, the conference record processing device 600 includes:

[0287] The optimal person determining module 601 is configured to determine, based on the current meeting record, a similar record with the highest similarity from multiple historical meeting records; and determine the optimal person to handle the current meeting record based on the similar records;

[0288] Comprehensive weight determination module 602 is configured to determine, based on the sources of multiple current meeting records, a set of current meeting records corresponding to different sources; and obtain a comprehensive weight of all current meeting records based on the weights of the moderators corresponding to the current meeting records in the set of current meeting records;

[0289] Processing module 603 is used to determine the pending records corresponding to each optimal processing personnel based on the correspondence between the current meeting records and the optimal processing personnel; and send multiple pending records corresponding to the optimal processing personnel and the corresponding comprehensive weights to the optimal processing personnel, so that the optimal processing personnel processes the corresponding pending records based on the comprehensive weights.

[0290] In an optional embodiment, the system further includes a meeting record extraction module for:

[0291] Before determining the most similar record from multiple historical meeting records based on the current meeting record,

[0292] Performing voice-to-text processing on the preset current conference voice to obtain a current conference record corresponding to the voice source of the current conference voice;

[0293] A projection set is formed based on multiple preset current conference projection images, and text is extracted from one of the current conference projection images in the projection set to obtain the current conference record corresponding to the projection source of the projection set.

[0294] In an optional embodiment, the meeting record extraction module is used to:

[0295] Based on the pixel information of all current conference projection images, the pixel similarity between all current conference projection images is obtained;

[0296] Repeating the step of forming a projection set until there is no current conference projection image that is not included in the projection set, the step of forming a projection set includes:

[0297] Select a target image from the current conference projection images that are not included in the projection set;

[0298] One of the projection sets is formed based on the selected images in the current conference projection images that are not included in the projection set, except the target image, whose pixel similarity with the target image is greater than a preset pixel similarity threshold, and the target image.

[0299] In an optional embodiment, the meeting record extraction module is used to:

[0300] Extract text from all current conference projection images to obtain the corresponding text content;

[0301] Based on the text content, obtaining the text similarity between all corresponding current conference projection images;

[0302] Repeating the step of forming a projection set until there is no current conference projection image that is not included in the projection set, the step of forming a projection set includes:

[0303] Select a target image from the current conference projection images that are not included in the projection set;

[0304] One of the projection sets is formed based on the selected images in the current conference projection images that are not included in the projection set, except the target image, whose text similarity with the target image is greater than a preset text similarity threshold, and the target image.

[0305] In an optional embodiment, the optimal personnel determination module 601 is configured to:

[0306] Performing word segmentation processing on the current meeting record to obtain a corresponding current vocabulary, and performing word segmentation processing on the plurality of historical meeting records to obtain historical vocabulary corresponding to the historical meeting records;

[0307] determining a record similarity between the current meeting record and each of the historical meeting records based on the current vocabulary and the historical vocabulary corresponding to each of the historical meeting records;

[0308] The historical meeting record with the highest record similarity is determined as the similar record.

[0309] In an optional embodiment, the optimal personnel determination module 601 is configured to:

[0310] Taking the intersection of the current vocabulary and the historical vocabulary corresponding to the historical meeting records to obtain a vocabulary intersection;

[0311] Taking a union of the current vocabulary and the historical vocabulary corresponding to the historical meeting records to obtain a vocabulary union;

[0312] Based on the number of words in the vocabulary intersection and the number of words in the vocabulary union, the record similarity between the current meeting record and the historical meeting record is obtained.

[0313] In an optional embodiment, the optimal personnel determination module 601 is configured to:

[0314] Based on the similar records, corresponding historical processing information is obtained;

[0315] Determining, based on the historical processing information, a plurality of processing times corresponding to each historical processing person;

[0316] Based on the current time and the processing time, obtaining time differences corresponding to multiple processing times of each historical processing personnel, and determining a maximum time difference based on all time differences corresponding to all historical processing personnel;

[0317] Obtaining a processing weight corresponding to the historical processing personnel according to multiple time differences corresponding to the historical processing personnel, a maximum time difference, and a preset correction coefficient;

[0318] The historical processing personnel with the largest processing weight is determined as the optimal processing personnel.

[0319] In an optional implementation, the comprehensive weight determination module 602 is configured to:

[0320] The current conference record whose source is the screen projection source is determined as the current conference screen projection record, and the current conference record whose source is the voice source is determined as the current conference voice record;

[0321] Determining a first current conference record set corresponding to a screen projection source according to the current conference screen projection record;

[0322] A second current conference record set corresponding to the voice source is determined based on the current conference voice record.

[0323] In an optional implementation, the comprehensive weight determination module 602 is configured to:

[0324] Perform word segmentation on each of the current conference screen projection records to obtain corresponding screen projection vocabulary;

[0325] Determining a first current similarity between all current conference screen projection records based on the screen projection vocabulary corresponding to the current conference screen projection record;

[0326] Repeating the step of forming a first current conference record set until there is no current conference screen projection record that is not included in the first current conference record set, the step of forming the first current conference record set includes:

[0327] Selecting a first target record from the current conference projection records that are not included in the first current conference record set;

[0328] Based on the first candidate record and the first target record in the current conference projection records that are not included in the first current conference record set, the first current similarity between the other records except the first target record and the first target record is greater than the preset current similarity threshold, one of the first current conference record sets is formed.

[0329] In an optional implementation, the comprehensive weight determination module 602 is configured to:

[0330] Intersections of two screen projection words corresponding to all the current conference screen projection records are taken to obtain a plurality of corresponding first intersections;

[0331] Taking the union of two screen projection words corresponding to all the current conference screen projection records to obtain multiple corresponding first unions;

[0332] Based on the number of words in the first intersection and the number of words in the first union, a first current similarity between the corresponding two current conference projection records is obtained.

[0333] In an optional implementation, the comprehensive weight determination module 602 is configured to:

[0334] Perform word segmentation processing on each of the current conference voice records to obtain corresponding voice vocabulary;

[0335] determining a second current similarity between all current conference voice records based on the voice vocabulary corresponding to the current conference voice record;

[0336] Repeating the step of forming a second current conference record set until there is no current conference voice record that is not included in the second current conference record set, the step of forming the second current conference record set includes:

[0337] Selecting a second target record from the current conference voice records that are not included in the second current conference record set;

[0338] One of the second current meeting record sets is formed based on the second candidate records and the second target record, whose second current similarity with the second target record in other records of the current meeting voice records other than the second target record is greater than the preset current similarity threshold.

[0339] In an optional implementation, the comprehensive weight determination module 602 is configured to:

[0340] Intersections of two voice words corresponding to all the current conference voice records are taken to obtain a plurality of corresponding second intersections;

[0341] Taking unions of two voice words corresponding to all the voice records of the current conference to obtain a plurality of corresponding second unions;

[0342] Based on the number of words in the second intersection and the number of words in the second union, a second current similarity between the corresponding two current conference voice records is obtained.

[0343] In an optional implementation, the comprehensive weight determination module 602 is configured to:

[0344] Determine whether the source corresponding to the current conference record set is a screen projection source or a voice source,

[0345] If it is a screen projection source, based on the current meeting record, obtain the number of projections of the current meeting projection image included in the corresponding projection set; based on the number of projections corresponding to the current meeting record and the host weight, obtain the corresponding sub-weight; superimpose the sub-weights corresponding to all current meeting records in the current meeting record set to obtain the projection sub-weight corresponding to the current meeting record set, and use the projection sub-weight as the projection sub-weight of each current meeting record in the current meeting record set;

[0346] If the source is voice, the host weights corresponding to all current meeting records in the current meeting record set are superimposed to obtain the voice sub-weight corresponding to the current meeting record set, and the voice sub-weight is used as the voice sub-weight of each current meeting record in the current meeting record set;

[0347] Based on the screen projection weights or voice weights corresponding to all current meeting records, the corresponding comprehensive weight is determined.

[0348] In an optional implementation, the comprehensive weight determination module 602 is configured to:

[0349] Determine whether the current conference record is only a screen projection record of the current conference, only a voice record of the current conference, or both a screen projection record and a voice record of the current conference;

[0350] If only the current conference screen projection record is used, the corresponding comprehensive weight is determined based on the screen projection weight and the preset screen projection coefficient;

[0351] If only the current conference voice record is used, the corresponding comprehensive weight is determined based on the voice sub-weight and the preset voice coefficient;

[0352] If it is both a screen projection record and a voice record of the current meeting, the corresponding first comprehensive weight is determined based on the screen projection weight and the preset screen projection coefficient, and the corresponding second comprehensive weight is determined based on the voice weight and the preset voice coefficient. The first comprehensive weight and the second comprehensive weight are added together to obtain the comprehensive weight.

[0353] In an optional implementation, the processing module 603 is configured to:

[0354] Based on the correspondence between each current meeting record in the current meeting record set and the optimal processing personnel, determining the number of current meeting records in the current meeting record set corresponding to different optimal processing personnel, and taking the optimal processing personnel corresponding to the largest number of current meeting records as the optimal processing personnel corresponding to the current meeting record set;

[0355] Selecting a representative record from each current meeting record set, and determining a correspondence between the optimal processing person and the representative record based on a correspondence between the representative record and the current meeting record set and a correspondence between the current meeting record set and the optimal processing person;

[0356] According to the correspondence between the optimal processing personnel and the representative records, all the representative records corresponding to the optimal processing personnel are determined as the records to be processed corresponding to the optimal processing personnel.

[0357] In an optional implementation, the processing module 603 is configured to:

[0358] After sorting the multiple pending records and the corresponding comprehensive weights corresponding to the optimal processing personnel according to the comprehensive weights, the records are sent to the optimal processing personnel so that the optimal processing personnel processes the corresponding pending records based on the sorted comprehensive weights.

[0359] Since the principle of solving the problem by the conference record processing device 600 is similar to that of the above method, the implementation of the conference record processing device 600 can refer to the implementation of the above method, which will not be repeated here.

[0360] The systems, devices, modules, or units described in the above embodiments may be implemented by computer chips or entities, or by products having certain functions. A typical implementation device is a computer device. Specifically, the computer device may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0361] In a typical example, a computer device specifically includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method described above is implemented.

[0362] Reference below Figure 7 , which shows a schematic structural diagram of a computer device 700 suitable for implementing an embodiment of the present application.

[0363] like Figure 7As shown, computer device 700 includes a central processing unit (CPU) 701, which can perform various appropriate tasks and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage portion 708 into a random access memory (RAM) 703. Various programs and data required for the operation of system 700 are also stored in RAM 703. CPU 701, ROM 702, and RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to bus 704.

[0364] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, and the like; an output section 707 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 708 including devices such as a hard disk; and a communication section 709 including a network interface card such as a LAN card or a modem. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. Removable media 711, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 710 as needed, so that computer programs read therefrom can be installed in the storage section 708 as needed.

[0365] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program including program code for executing the methods illustrated in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 709 and / or installed from removable media 711.

[0366] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0367] For the convenience of description, the above devices are described as being divided into various units according to their functions. Of course, when implementing this application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0368] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0369] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0370] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0371] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "includes a..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0372] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0373] The present application may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communications network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0374] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0375] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method for processing conference records, characterized in that: include: Performing voice-to-text processing on the preset current conference voice to obtain a current conference record corresponding to the voice source of the current conference voice; According to multiple preset current conference projection images, a projection set is formed, including: based on the pixel information of all current conference projection images, obtaining the pixel similarity between all current conference projection images; repeatedly performing the step of forming the projection set until there are no current conference projection images that are not included in the projection set, the step of forming the projection set including: selecting a target image from the current conference projection images that are not included in the projection set; based on the selected images in the current conference projection images that are not included in the projection set, whose pixel similarity with the target image is greater than a preset pixel similarity threshold and the target image, and the target image, forming one of the projection sets; Extract text from one of the current conference projection images in the projection set to obtain a current conference record corresponding to the projection source of the projection set; According to the current meeting record, determine the similar record with the highest similarity from multiple historical meeting records; based on the similar records, determine the optimal person to handle the current meeting record; Determining, based on the sources of multiple current meeting records, current meeting record sets corresponding to different sources; and obtaining a comprehensive weight of all current meeting records based on the weights of the moderators corresponding to the current meeting records in the current meeting record set; According to the correspondence between the current meeting record and the optimal processing personnel, the pending records corresponding to each optimal processing personnel are determined; the multiple pending records corresponding to the optimal processing personnel and the corresponding comprehensive weights are sent to the optimal processing personnel, so that the optimal processing personnel processes the corresponding pending records based on the comprehensive weights.

2. The method according to claim 1, characterized in that The forming of a projection set based on a plurality of preset current conference projection images includes: Extract text from all current conference projection images to obtain the corresponding text content; Based on the text content, obtaining the text similarity between all corresponding current conference projection images; Repeating the step of forming a projection set until there is no current conference projection image that is not included in the projection set, the step of forming a projection set includes: Select a target image from the current conference projection images that are not included in the projection set; One of the projection sets is formed based on the selected images in the current conference projection images that are not included in the projection set, except the target image, whose text similarity with the target image is greater than a preset text similarity threshold, and the target image.

3. The method according to claim 1, characterized in that The method of determining the similar record with the highest similarity from multiple historical meeting records based on the current meeting record includes: Performing word segmentation processing on the current meeting record to obtain a corresponding current vocabulary, and performing word segmentation processing on the plurality of historical meeting records to obtain historical vocabulary corresponding to the historical meeting records; determining a record similarity between the current meeting record and each of the historical meeting records based on the current vocabulary and the historical vocabulary corresponding to each of the historical meeting records; The historical meeting record with the highest record similarity is determined as the similar record.

4. The method according to claim 3, characterized in that The determining, based on the current vocabulary and the historical vocabulary corresponding to each of the historical meeting records, the record similarity between the current meeting record and each of the historical meeting records includes: Taking the intersection of the current vocabulary and the historical vocabulary corresponding to the historical meeting records to obtain a vocabulary intersection; Taking a union of the current vocabulary and the historical vocabulary corresponding to the historical meeting records to obtain a vocabulary union; Based on the number of words in the vocabulary intersection and the number of words in the vocabulary union, the record similarity between the current meeting record and the historical meeting record is obtained.

5. The method according to claim 1, characterized in that Determining the optimal person to handle the current meeting record based on the similar records includes: Based on the similar records, corresponding historical processing information is obtained; Determining, based on the historical processing information, a plurality of processing times corresponding to each historical processing person; Based on the current time and the processing time, obtaining time differences corresponding to multiple processing times of each historical processing personnel, and determining a maximum time difference based on all time differences corresponding to all historical processing personnel; Obtaining a processing weight corresponding to the historical processing personnel according to multiple time differences corresponding to the historical processing personnel, a maximum time difference, and a preset correction coefficient; The historical processing personnel with the largest processing weight is determined as the optimal processing personnel.

6. The method according to claim 1, characterized in that The step of determining, based on the sources of the plurality of current meeting records, current meeting record sets corresponding to different sources respectively includes: The current conference record whose source is the screen projection source is determined as the current conference screen projection record, and the current conference record whose source is the voice source is determined as the current conference voice record; Determining a first current conference record set corresponding to a screen projection source according to the current conference screen projection record; A second current conference record set corresponding to the voice source is determined based on the current conference voice record.

7. The method according to claim 6, characterized in that The determining, based on the current conference screen projection record, a first current conference record set corresponding to the screen projection source includes: Perform word segmentation on each of the current conference screen projection records to obtain corresponding screen projection vocabulary; Determining a first current similarity between all current conference screen projection records based on the screen projection vocabulary corresponding to the current conference screen projection record; Repeating the step of forming a first current conference record set until there is no current conference screen projection record that is not included in the first current conference record set, the step of forming the first current conference record set includes: Selecting a first target record from the current conference projection records that are not included in the first current conference record set; Based on the first candidate record and the first target record in the current conference projection records that are not included in the first current conference record set, the first current similarity between the other records except the first target record and the first target record is greater than the preset current similarity threshold, one of the first current conference record sets is formed.

8. The method according to claim 7, characterized in that The determining, based on the screen projection vocabulary corresponding to the current conference screen projection record, a first current similarity between all current conference screen projection records includes: Intersections of two screen projection words corresponding to all the current conference screen projection records are taken to obtain a plurality of corresponding first intersections; Taking the union of two screen projection words corresponding to all the current conference screen projection records to obtain multiple corresponding first unions; Based on the number of words in the first intersection and the number of words in the first union, a first current similarity between the corresponding two current conference projection records is obtained.

9. The method according to claim 6, characterized in that The determining, based on the current conference voice record, a second current conference record set corresponding to the voice source includes: Perform word segmentation processing on each of the current conference voice records to obtain corresponding voice vocabulary; determining a second current similarity between all current conference voice records based on the voice vocabulary corresponding to the current conference voice record; Repeating the step of forming a second current conference record set until there is no current conference voice record that is not included in the second current conference record set, the step of forming the second current conference record set includes: Selecting a second target record from the current conference voice records that are not included in the second current conference record set; One of the second current meeting record sets is formed based on the second candidate records and the second target record, whose second current similarity with the second target record in other records of the current meeting voice records other than the second target record is greater than the preset current similarity threshold.

10. The method according to claim 9, characterized in that The determining, based on the voice vocabulary corresponding to the current conference voice record, a second current similarity between all current conference voice records includes: Intersections of two voice words corresponding to all the current conference voice records are taken to obtain a plurality of corresponding second intersections; Taking unions of two voice words corresponding to all the voice records of the current conference to obtain a plurality of corresponding second unions; Based on the number of words in the second intersection and the number of words in the second union, a second current similarity between the corresponding two current conference voice records is obtained.

11. The method according to claim 6, characterized in that The method of obtaining the comprehensive weight of all current meeting records based on the weight of the host corresponding to the current meeting record in the current meeting record set includes: Determine whether the source corresponding to the current conference record set is a screen projection source or a voice source, If it is a screen projection source, based on the current meeting record, obtain the number of projections of the current meeting projection image included in the corresponding projection set; based on the number of projections corresponding to the current meeting record and the host weight, obtain the corresponding sub-weight; superimpose the sub-weights corresponding to all current meeting records in the current meeting record set to obtain the projection sub-weight corresponding to the current meeting record set, and use the projection sub-weight as the projection sub-weight of each current meeting record in the current meeting record set; If the source is voice, the host weights corresponding to all current meeting records in the current meeting record set are superimposed to obtain the voice sub-weight corresponding to the current meeting record set, and the voice sub-weight is used as the voice sub-weight of each current meeting record in the current meeting record set; Based on the screen projection weights or voice weights corresponding to all current meeting records, the corresponding comprehensive weight is determined.

12. The method according to claim 11, characterized in that The determining of the corresponding comprehensive weight based on the screen projection weights or voice weights corresponding to all current conference records includes: Determine whether the current conference record is only a screen projection record of the current conference, only a voice record of the current conference, or both a screen projection record and a voice record of the current conference; If only the current conference screen projection record is used, the corresponding comprehensive weight is determined based on the screen projection weight and the preset screen projection coefficient; If only the current conference voice record is used, the corresponding comprehensive weight is determined based on the voice sub-weight and the preset voice coefficient; If it is both a screen projection record and a voice record of the current meeting, the corresponding first comprehensive weight is determined based on the screen projection weight and the preset screen projection coefficient, and the corresponding second comprehensive weight is determined based on the voice weight and the preset voice coefficient. The first comprehensive weight and the second comprehensive weight are added together to obtain the comprehensive weight.

13. The method according to claim 1, wherein The step of determining the to-be-processed record corresponding to each optimal processing personnel according to the correspondence between the current meeting record and the optimal processing personnel includes: Based on the correspondence between each current meeting record in the current meeting record set and the optimal processing personnel, determining the number of current meeting records in the current meeting record set corresponding to different optimal processing personnel, and taking the optimal processing personnel corresponding to the largest number of current meeting records as the optimal processing personnel corresponding to the current meeting record set; Selecting a representative record from each current meeting record set, and determining a correspondence between the optimal processing person and the representative record based on a correspondence between the representative record and the current meeting record set and a correspondence between the current meeting record set and the optimal processing person; According to the correspondence between the optimal processing personnel and the representative records, all the representative records corresponding to the optimal processing personnel are determined as the records to be processed corresponding to the optimal processing personnel.

14. The method according to claim 1, wherein The step of sending the plurality of pending records corresponding to the optimal processing personnel and the corresponding comprehensive weights to the optimal processing personnel so that the optimal processing personnel processes the corresponding pending records based on the comprehensive weights includes: After sorting the multiple pending records and the corresponding comprehensive weights corresponding to the optimal processing personnel according to the comprehensive weights, the records are sent to the optimal processing personnel so that the optimal processing personnel processes the corresponding pending records based on the sorted comprehensive weights.

15. A conference record processing device, characterized in that: include: A conference record extraction module is used to perform speech-to-text processing on the preset current conference voice to obtain the current conference record corresponding to the voice source of the current conference voice; According to multiple preset current conference projection images, a projection set is formed, including: based on the pixel information of all current conference projection images, obtaining the pixel similarity between all current conference projection images; repeatedly performing the step of forming the projection set until there are no current conference projection images that are not included in the projection set, the step of forming the projection set including: selecting a target image from the current conference projection images that are not included in the projection set; based on the selected images in the current conference projection images that are not included in the projection set, whose pixel similarity with the target image is greater than a preset pixel similarity threshold and the target image, and the target image, forming one of the projection sets; Extract text from one of the current conference projection images in the projection set to obtain a current conference record corresponding to the projection source of the projection set; An optimal personnel determination module is used to determine a similar record with the highest similarity from multiple historical meeting records based on the current meeting record; and based on the similar records, determine the optimal person to handle the current meeting record; A comprehensive weight determination module is used to determine, based on the sources of multiple current meeting records, a set of current meeting records corresponding to different sources; and obtain a comprehensive weight of all current meeting records based on the weights of the host corresponding to the current meeting records in the set of current meeting records; A processing module is used to determine the pending records corresponding to each optimal processing personnel based on the correspondence between the current meeting records and the optimal processing personnel; and send multiple pending records corresponding to the optimal processing personnel and the corresponding comprehensive weights to the optimal processing personnel, so that the optimal processing personnel processes the corresponding pending records based on the comprehensive weights.

16. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method according to any one of claims 1 to 14 is implemented.

17. A computer-readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 14 is implemented.

Images