Intelligent-based talent resume screening and matching method and device

Abstract

The invention provides a talent resume screening and matching method and device based on intelligence. The method comprises the following steps: carrying out region extraction on resume files sent by different talent ends to obtain resume regions corresponding to different universal attributes and configured on the basis of a management end; performing resume evaluation based on a resume area in each resume file, and performing resume ranking on each resume file from high to low based on each resume screening score to obtain a resume sequence; performing resume screening on the resume sequence, determining each obtained resume file corresponding to a preset screening proportion as a resume screening group, calling a resume template configured by the management end, and sending the resume template to each talent end corresponding to each resume file in the resume screening group for filling; and receiving a talent resume uploaded after any talent end fills in the resume template, triggering the interview end to carry out video interview on the talent end, and carrying out analysis processing on the video interview to obtain an interview score of the talent end. The resume screening efficiency is at least improved.

Description

Technical Field

[0001] The present invention relates to data processing technology, and in particular to a talent resume screening and matching method and device based on intelligence. Background Art

[0002] In the recruitment process, screening of talent resumes is an essential step. Through resume screening, companies can compare candidates with job requirements and select matching candidates.

[0003] The inventors found in their research that existing resume screening methods are mostly done manually, so there is a possibility of incorrect matches during resume screening, and manual resume screening takes a lot of time, which may extend the recruitment cycle and increase the corresponding costs. Summary of the invention

[0004] Based on the above problems, the present invention is proposed to provide a method and device for screening and matching talent resumes based on intelligence, which overcomes the above problems or at least partially solves the above problems.

[0005] According to one aspect of the present invention, there is provided a talent resume screening and matching method based on intelligence, comprising the following steps: extracting regions from each resume file sent by each different talent terminal to obtain resume regions corresponding to different general attributes based on the configuration of the management terminal; Resume evaluation is performed based on resume areas corresponding to different general attributes respectively located in each resume file, and resume ranking of each resume file is performed from high to low based on the obtained resume screening scores to obtain a resume sequence; Screening the resume sequence based on a preset screening ratio, determining the obtained resume files corresponding to the preset screening ratio as a resume screening group, retrieving a resume template including filling areas corresponding to different unique attributes configured based on the management end, and sending the template to each talent end corresponding to each resume file in the resume screening group for filling; In response to receiving a talent resume uploaded by any talent after filling in the resume template, the interview end is triggered to conduct a video interview with the talent end based on the filling areas corresponding to different unique attributes in the talent resume, and analyze and process the video interview based on a preset interview analysis strategy to obtain an interview score corresponding to the talent end.

[0006] Optionally, in the method according to the present invention, region extraction is performed on each resume file sent by each different talent terminal to obtain resume regions corresponding to different general attributes configured by the management terminal, including: The structures of the resume files sent by different talent terminals are obtained respectively, and the file segments in the same resume file are obtained, and the resume regions formed by the adjacent file segments are determined as region extraction groups; Perform pixel identification on each resume area in the area extraction group, and remove all resume areas that do not contain any pixel from the area extraction group to obtain an updated area extraction group; Performing character recognition on each resume region in the updated region extraction group, and obtaining common key contents corresponding to different common attributes configured by the management end; Each resume area is traversed, and a resume area including any of the universal key contents is determined as a universal attribute corresponding to the universal key content, thereby obtaining resume areas corresponding to different universal attributes.

[0007] Optionally, in the method according to the present invention, each resume area formed by each adjacent file line segment is determined as an area extraction group, including: Performing coordinate processing on each resume file respectively, and determining each area coordinate group corresponding to each resume area based on each resume file after coordinate processing; The regional coordinate groups corresponding to the coordinate extreme values ​​in each regional coordinate group are removed, and all the remaining regional coordinate groups are compared in pairs, and it is determined whether there is an enclosing relationship between the regional coordinate groups based on the comparison results; The region coordinate groups having an enclosing relationship are merged into a resume region, the region coordinate groups not having an enclosing relationship are respectively determined as different resume regions, and the obtained resume regions are determined as region extraction groups.

[0008] Optionally, in the method according to the present invention, resume evaluation is performed based on resume areas corresponding to different general attributes respectively located in each resume file, and resume ranking of each resume file is performed from high to low based on the obtained resume screening scores to obtain a resume sequence, including: Obtaining general screening conditions corresponding to different general attributes configured on the management side, and obtaining general area semantics corresponding to the resume areas corresponding to different general attributes; Based on any general screening condition, a regional evaluation is performed on the general regional semantics corresponding to the same general attribute, and based on the evaluation result, a score of each attribute corresponding to each general attribute is determined; Based on the average calculation of each attribute score, each resume screening score corresponding to each resume file is obtained, and based on the obtained resume screening score, each resume file is ranked from high to low to obtain a resume sequence, wherein the resume sequence includes each resume file and a ranking number corresponding to each resume file.

[0009] Optionally, in the method according to the present invention, based on any general screening condition, a general regional semantics corresponding to the same general attribute is evaluated, and based on the evaluation result, the attribute scores corresponding to each general attribute are determined, including: When any of the common attributes include numeric attributes, determining a numerical filtering interval based on a general filtering condition corresponding to the numerical attribute, determining a region value based on a general region semantics corresponding to the numerical attribute, and comparing the region value with the numerical filtering interval; When the region value is within the value screening interval, determining the attribute score corresponding to the general attribute as the first value score; When the region value is outside the value screening interval, a maximum value and a minimum value in the value screening interval are determined, and a maximum difference and a minimum difference between the region value and the maximum value and the minimum value are determined respectively; The difference with the smallest corresponding value is determined as the evaluation difference, and the evaluation difference is multiplied by the retrieved difference coefficient to obtain the evaluation elimination value; Calculating the difference between the first numerical score and the evaluation elimination value, and determining the obtained second numerical score as the attribute score corresponding to the general attribute; and / or When any common attribute includes a noun attribute, Determining noun screening content based on the general screening conditions corresponding to the noun attributes, determining regional nouns based on the general regional semantics corresponding to the noun attributes, and comparing the regional nouns with the noun screening content; When the regional noun has a matching relationship with the noun screening content, determining the attribute score corresponding to the general attribute as the first noun score; When the regional noun does not have a matching relationship with the noun screening content, the attribute score corresponding to the general attribute is determined as the second noun score.

[0010] Optionally, in the method according to the present invention, the resume sequence is screened based on a preset screening ratio, each resume file corresponding to the preset screening ratio is determined as a resume screening group, and a resume template including filling areas corresponding to different unique attributes configured based on the management end is retrieved and sent to each talent end corresponding to each resume file in the resume screening group for filling, including: Retrieving historical resume feedback data, wherein the historical resume feedback data includes different historical periods and the resume feedback ratios corresponding to the different historical periods; Calculate the average of each resume feedback ratio to obtain the mean ratio; Retrieving a first preset demand quantity, and calculating the ratio of the first preset demand quantity to the mean of the ratio to obtain a second preset demand quantity; Based on the second preset required number, the resume sequence is screened from high to low, and each resume file corresponding to the second preset required number is determined as a resume screening group; Retrieving a resume template configured based on the management end, wherein the resume template includes filling areas corresponding to different unique attributes; Based on the resume template, each resume file in the resume screening group is configured one by one, and each filling area in the resume template is updated based on each resume file; Send the resume templates that have been updated by the region to the corresponding talent terminals for filling out.

[0011] Optionally, in the method according to the present invention, each filling area located in the resume template is updated based on each resume file, including: All remaining areas except the resume area corresponding to the general attribute in each resume file are respectively determined as the areas to be matched, and the semantic labels to be filled in corresponding to the filled-in areas are obtained; Performing semantic recognition on each to-be-matched region in each resume file to obtain the semantics of each to-be-matched region corresponding to each to-be-matched region; Match each to-be-matched region semantic with the to-be-filled semantic label one by one, and determine the to-be-matched region semantic that has a matching relationship with any to-be-filled semantic label as the matched region semantic; Based on the matching relationship, the to-be-matched area corresponding to the semantics of the matched area is mapped to the filling area corresponding to the semantic label to be filled, and all the mapped filling areas are converted from an initial to-be-filled state to a filled state; The area outlines of all filled-in areas in the resume template are configured with a first color, and the area outlines of all filled-in areas to be filled-in are configured with a second color different from the first color.

[0012] Optionally, in the method according to the present invention, in response to receiving a resume uploaded by any talent terminal after filling in the resume template, triggering the interview terminal to conduct a video interview with the talent terminal based on the filling areas corresponding to different unique attributes in the talent resume, and analyzing and processing the video interview based on a preset interview analysis strategy to obtain an interview score corresponding to the talent terminal, including: Create an interview record layer, and send the interview record layer to the interview end, wherein the interview record layer includes a resume display area, a video display area, and a key point record area; In response to the interview end establishing a video call with the talent end, mapping the video call to the video display area, and mapping the filling areas corresponding to different unique attributes in the talent resume to the resume display area for display; In response to the interaction of the interview end with any filling area, the state acquisition unit is triggered to collect the state of the video portion of the corresponding talent end located in the video display area, obtain the interview state corresponding to the talent end, and establish an active recording sub-area and a passive recording sub-area corresponding to the filling area in the key point recording area; Acquire the interview content fed back by the talent end based on the filling area, compare the interview content with the filling content in the filling area, and determine the subjective evaluation content corresponding to the filling area based on the comparison result; Adding the subjective evaluation content to the active recording sub-area, adding the interview status to the passive recording sub-area, and determining the active recording sub-area and the passive recording sub-area corresponding to the same filling area after the addition as a key content group; In response to completing the determination of the key content groups corresponding to each of the filled-in areas, performing content evaluation on each of the key content groups to obtain original scores corresponding to each of the key content groups; Obtain the scoring weights corresponding to the filled-in areas, and multiply the original scores by the corresponding scoring weights to obtain the scores for each key point; The scores of each key point are averaged, and the obtained average score is determined as the interview score corresponding to the talent end.

[0013] Optionally, in the method according to the present invention, the triggering state acquisition unit performs state acquisition on the video portion of the corresponding talent terminal located in the video display area to obtain the interview state corresponding to the talent terminal, including: The trigger state acquisition unit performs face recognition on the video portion to obtain a face recognition area corresponding to the talent end, and performs expression recognition and voice recognition based on the face recognition area to obtain an expression recognition state and a voice recognition state corresponding to the talent end; Retrieving a preset expression evaluation table, wherein the preset expression evaluation table includes expression evaluation scores corresponding to different expression states; Traversing the preset expression evaluation table to determine the expression evaluation score corresponding to the expression recognition state; Retrieving a preset sound evaluation table, wherein the preset sound evaluation table includes sound evaluation scores corresponding to different sound states; Traversing the preset sound evaluation table to determine the sound evaluation score corresponding to the sound recognition state; Based on the expression evaluation score and the voice evaluation score, a preset proportion is summed, and based on each summation result, an emotion score corresponding to each staff member is determined; Retrieving a status evaluation table, wherein the status evaluation table includes different score intervals and different interview states corresponding to the different score intervals; The state evaluation table is traversed to determine the interview state corresponding to the score interval containing the emotion score.

[0014] Optionally, in the method according to the present invention, obtaining the interview content fed back by the talent end based on the filling area, comparing the interview content with the filling content in the filling area, and determining the subjective evaluation content corresponding to the filling area based on the comparison result, includes: Acquire the interview content fed back by the talent end based on the filling area, and perform semantic recognition on the interview content and the filling content respectively to obtain the interview semantics and the filling semantics corresponding to the interview content and the filling content respectively; The interview semantics is semantically compared with the filled-in semantics, and the subjective evaluation content corresponding to the filled-in area is determined based on the comparison result, wherein the subjective evaluation content includes positive evaluation and negative evaluation.

[0015] Optionally, in the method according to the present invention, in response to completing the determination of the key content groups corresponding to each fill-in area, performing content evaluation on each key content group to obtain original scores corresponding to each key content group includes: In response to completing the determination of the key content groups corresponding to each filling area, obtaining the subjective evaluation content and the interview status respectively included in each key content group; In response to the subjective evaluation content included in any key content group being a negative evaluation, determining the retrieved first preset score as an original score corresponding to the key content group; In response to the subjective evaluation content included in any key content group being a positive evaluation, determining the retrieved second preset score as the subjective score corresponding to the subjective evaluation content, and retrieving a preset state table, wherein the preset state table includes different interview states and state scores corresponding to the different interview states; Obtaining an interview state that is in the same key content group as the subjective evaluation content, and traversing the preset state table to determine a state score corresponding to the interview state; The second preset score and the state score are summed up and the sum result is determined as the original score corresponding to the key content group.

[0016] According to another aspect of the present invention, there is provided a talent resume screening and matching device based on intelligence, comprising: The region extraction module is configured to perform region extraction on each resume file sent by each different talent terminal, and obtain the resume regions corresponding to different general attributes configured by the management terminal; A sequence generation module is configured to evaluate resumes based on resume areas corresponding to different general attributes in each resume file, and to rank the resume files from high to low based on the obtained resume screening scores to obtain a resume sequence; A resume screening module is configured to screen the resume sequence based on a preset screening ratio, determine the obtained resume files corresponding to the preset screening ratio as a resume screening group, retrieve a resume template including filling areas corresponding to different unique attributes configured based on the management end, and send it to each talent end corresponding to each resume file in the resume screening group for filling; The interview scoring module is configured to respond to the receipt of a talent resume uploaded by any talent end after filling in the resume template, trigger the interview end to conduct a video interview with the talent end based on the filling areas corresponding to different unique attributes in the talent resume, and analyze and process the video interview based on a preset interview analysis strategy to obtain an interview score corresponding to the talent end.

[0017] According to the solution of the present invention, the server will first extract the various different areas such as the personal photo area in each resume file sent by each different talent terminal, so as to obtain the resume areas corresponding to different general attributes configured by the management terminal, so as to facilitate the subsequent management terminal to screen each resume file. The server will also evaluate the resume area corresponding to the general attribute, and then obtain the resume screening score corresponding to the different resume files according to each resume evaluation value. Then, the server will rank each resume file in the order of high to low resume screening scores to obtain a resume sequence, so that the subsequent management terminal can check which resumes are high-quality resumes according to the resume sequence. Then, the server will perform a preliminary screening of the resume sequence based on the preset screening ratio, so as to determine the obtained resume files corresponding to the preset screening ratio as a resume screening group. And the management terminal can configure a resume template containing each filling area corresponding to different unique attributes according to the job recruitment situation. The server will retrieve the resume template and send it to each talent terminal corresponding to each resume file in the resume screening group for filling. When any talent end completes filling in the resume template, the corresponding talent resume will be obtained. When the server receives the talent resume uploaded by the talent end, it will trigger the interview end to conduct a video interview with the talent end according to the filling areas corresponding to different unique attributes in the talent resume, and analyze and process this video interview based on a preset interview analysis strategy, so as to obtain an interview score corresponding to the talent end. The present invention can improve the efficiency of resume screening, and obtain an interview score for the video interview, which can provide a certain reference to the management end. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flow chart of a method for screening and matching talent resumes based on intelligence according to an embodiment of the present invention is shown; Figure 2 A schematic diagram of a region extraction group according to an embodiment of the present invention is shown; Figure 3 A schematic diagram showing an active recording sub-area and a passive recording sub-area according to an embodiment of the present invention is shown; Figure 4 A structural block diagram of an intelligent talent resume screening and matching device according to another embodiment of the present invention is shown. DETAILED DESCRIPTION

[0019] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0020] In the recruitment process, screening of talent resumes is an essential step. Through resume screening, companies can compare candidates with job requirements and select matching candidates.

[0021] The inventors found in their research that existing resume screening methods are mostly done manually, so there is a possibility of incorrect matches during resume screening, and manual resume screening takes a lot of time, which may extend the recruitment cycle and increase the corresponding costs.

[0022] To solve the problems existing in the above-mentioned prior art, the inventor proposes the solution of the present invention. An embodiment of the present invention provides a method for screening and matching talent resumes based on intelligence, which can be executed in a computing device, wherein the computing device can be understood as a terminal with data processing function, such as a mobile phone or a server.

[0023] Figure 1 A flow chart of a talent resume screening and matching method based on intelligence according to an embodiment of the present invention is shown, and the method is suitable for being executed in a computing device.

[0024] like Figure 1 As shown, the purpose of this embodiment is to implement a talent resume screening and matching method based on intelligence, starting from step S102, in step S102, including the following steps: Perform regional extraction on each resume file sent by different talent terminals to obtain resume regions corresponding to different general attributes based on the configuration of the management terminal.

[0025] For example, in this embodiment, the talent end can be understood as a candidate participating in the application. Since the resume files sent by different talent ends are made according to their own styles, the content layout of each resume file may be different. For example, candidate A will set his personal photo in the area located in the upper right corner of the resume file, while candidate B will set the personal photo area in the area located in the upper left corner of the resume file, etc.

[0026] And because different positions have different job requirements during recruitment, such as age requirements or graduation school requirements, there will be different general attributes in the resume file. For example, when the recruitment position has age restrictions on employees, age is a general attribute. Therefore, the server will first extract the various different areas such as the personal photo area, personal profile area or work experience area in each resume file sent by different talent terminals, that is, perform area extraction, and obtain the resume areas corresponding to different general attributes based on the configuration of the management terminal, such as the resume area corresponding to age, so that the subsequent management terminal can conduct preliminary screening of each resume file based on various general attributes.

[0027] Furthermore, the above-mentioned “extracting regions from the resume files sent by the different talent terminals to obtain resume regions corresponding to different general attributes configured by the management terminal” further includes the following steps: The structures of the resume files sent by different talent terminals are obtained respectively, and the file segments in the same resume file are obtained, and the resume regions formed by the adjacent file segments are determined as region extraction groups; Perform pixel identification on each resume area in the area extraction group, and remove all resume areas that do not contain any pixel from the area extraction group to obtain an updated area extraction group; Performing character recognition on each resume region in the updated region extraction group, and obtaining common key contents corresponding to different common attributes configured by the management end; Each resume area is traversed, and a resume area including any of the universal key contents is determined as a universal attribute corresponding to the universal key content, thereby obtaining resume areas corresponding to different universal attributes.

[0028] For example, in this embodiment, since each resume area in the resume file is surrounded by a rectangular frame, there will be many horizontal and vertical lines in the resume file, that is, each file segment. Therefore, in order to determine each resume area separately, the server will first obtain the structure of each resume file sent by each different talent terminal, thereby obtaining each file segment in the same resume file, and determine each resume area composed of adjacent file segments as a region extraction group, such as Figure 2 shown.

[0029] When candidates are making resumes, some resume areas may not be filled with content. Therefore, the server will perform pixel recognition on each resume area in the area extraction group. When it is recognized that a resume area does not contain pixels, it means that the resume area is blank. Therefore, the server will remove all resume areas that do not contain any pixels from the area extraction group, thereby obtaining an updated area extraction group.

[0030] Then, the server will perform character recognition on each resume area in the updated area extraction group, thereby obtaining the common key content corresponding to different common attributes based on the configuration of the management end. For example, if the common attribute is age, the obtained common key content may be "Age: 30 years old". Next, the server will traverse each resume area and determine the resume area including the common key content as the common attribute corresponding to the common key content, thereby obtaining resume areas corresponding to different common attributes.

[0031] This embodiment can remove blank resume areas by first performing pixel recognition on each resume area in the area extraction group when performing area extraction, thereby obtaining an updated area extraction group, which can reduce the subsequent workload of the server to a certain extent.

[0032] Furthermore, the above-mentioned "determining each resume area formed by each adjacent file line segment as an area extraction group" further includes the following steps: Performing coordinate processing on each resume file respectively, and determining each area coordinate group corresponding to each resume area based on each resume file after coordinate processing; The regional coordinate groups corresponding to the coordinate extreme values ​​in each regional coordinate group are removed, and all the remaining regional coordinate groups are compared in pairs, and it is determined whether there is an enclosing relationship between the regional coordinate groups based on the comparison results; The region coordinate groups having an enclosing relationship are merged into a resume region, the region coordinate groups not having an enclosing relationship are respectively determined as different resume regions, and the obtained resume regions are determined as region extraction groups.

[0033] For example, in this embodiment, when a candidate prepares a resume file, a resume area may be subdivided into multiple sub-areas. For example, a talent has two work experiences, so two sub-areas are divided in the resume area corresponding to the work experience, and the two work experiences are filled in the two sub-areas in sequence. At this time, a rectangular box contains multiple small rectangular boxes. Therefore, the server determines whether there are rectangular boxes surrounding each other in the resume file in the following way.

[0034] First, the server will coordinate each resume file and determine the coordinate groups of each resume area based on each resume file after coordinate processing. When creating a resume file on the talent side, a large rectangular frame that can contain all resume areas may be set for aesthetics. In order to avoid the influence of the large rectangular frame on subsequent steps, the server will determine whether such a large rectangular frame is included in each resume file.

[0035] Since each line segment constituting such a large rectangular frame is located at the edge of the resume file, the server can first determine the regional coordinate group with the corresponding coordinate extreme value in each regional coordinate group, and then remove the regional coordinate group. Then, the server will compare all the remaining regional coordinate groups in pairs, and determine whether there is an enclosing relationship between the regional coordinate groups according to the comparison results.

[0036] Next, the server will merge the regional coordinate groups with an enclosing relationship into a resume area, and determine the regional coordinate groups without an enclosing relationship as different resume areas, and determine the obtained resume areas as area extraction groups.

[0037] In step S104, the following contents are included: Resume evaluation is performed based on resume areas corresponding to different general attributes in each resume file, and resume ranking of each resume file is performed from high to low based on the obtained resume screening scores to obtain a resume sequence.

[0038] For example, in this embodiment, the company may have requirements for both the age and years of experience of the candidate, that is, there may be multiple common attributes. At this time, the server will evaluate the resume areas corresponding to the common attributes in turn, and then obtain the resume screening scores corresponding to different resume files according to the resume evaluation values. Then, the server will rank the resume files in descending order according to the resume screening scores, thereby obtaining a resume sequence. This allows the subsequent management end to view which resumes are high-quality resumes and which resumes are low-quality resumes based on the resume sequence.

[0039] Furthermore, the above-mentioned "performing resume evaluation based on the resume areas corresponding to different general attributes respectively located in each resume file, and ranking the resume files from high to low based on the obtained resume screening scores to obtain a resume sequence" also includes the following steps: Obtaining general screening conditions corresponding to different general attributes configured on the management side, and obtaining general area semantics corresponding to the resume areas corresponding to different general attributes; Based on any general screening condition, a regional evaluation is performed on the general regional semantics corresponding to the same general attribute, and based on the evaluation result, a score of each attribute corresponding to each general attribute is determined; Based on the average calculation of each attribute score, each resume screening score corresponding to each resume file is obtained, and based on the obtained resume screening score, each resume file is ranked from high to low to obtain a resume sequence, wherein the resume sequence includes each resume file and a ranking number corresponding to each resume file.

[0040] For example, in this embodiment, the server will first obtain the general screening conditions corresponding to different general attributes configured by the management end, for example, the general attribute is age, and the general screening condition is under 35 years old. Then, the server will obtain the regional semantics in the resume area corresponding to the general attribute of age, that is, obtain the general regional semantics corresponding to the resume areas corresponding to different general attributes.

[0041] Then, the server will perform regional evaluation on the regional semantics corresponding to the same general attribute according to the general screening conditions, thereby obtaining evaluation results, and then determine the attribute scores corresponding to each general attribute according to the evaluation results.

[0042] Since there may be multiple attribute scores, the server will calculate the average of each attribute score to obtain the resume screening score of each resume file, and then rank each resume file in descending order according to the resume screening score to obtain a resume sequence. The resume sequence includes each resume file and the ranking number corresponding to each resume file, which is convenient for the subsequent management end to view it accordingly.

[0043] Furthermore, the above-mentioned "based on any general screening condition, performing regional evaluation on the general regional semantics corresponding to the same general attribute, and determining the attribute scores corresponding to each general attribute respectively based on the evaluation result" also includes the following steps: When any of the common attributes include numeric attributes, determining a numerical filtering interval based on a general filtering condition corresponding to the numerical attribute, determining a region value based on a general region semantics corresponding to the numerical attribute, and comparing the region value with the numerical filtering interval; When the region value is within the value screening interval, determining the attribute score corresponding to the general attribute as the first value score; When the region value is outside the value screening interval, a maximum value and a minimum value in the value screening interval are determined, and a maximum difference and a minimum difference between the region value and the maximum value and the minimum value are determined respectively; The difference with the smallest corresponding value is determined as the evaluation difference, and the evaluation difference is multiplied by the retrieved difference coefficient to obtain the evaluation elimination value; Calculating the difference between the first numerical score and the evaluation elimination value, and determining the obtained second numerical score as the attribute score corresponding to the general attribute; and / or When any common attribute includes a noun attribute, Determining noun screening content based on the general screening conditions corresponding to the noun attributes, determining regional nouns based on the general regional semantics corresponding to the noun attributes, and comparing the regional nouns with the noun screening content; When the regional noun has a matching relationship with the noun screening content, determining the attribute score corresponding to the general attribute as the first noun score; When the regional noun does not have a matching relationship with the noun screening content, the attribute score corresponding to the general attribute is determined as the second noun score.

[0044] For example, in this embodiment, some general attributes can be represented by numerical values, such as the general screening condition of being under 35 years old or the school ranking in the top 100, etc., that is, the corresponding general attributes are general attributes including numerical attributes; some general attributes involve some specific nouns, such as the general screening condition of being a software engineer, etc., that is, the corresponding general attributes are general attributes including noun attributes. Therefore, the server will set different scoring methods for the two different types of general attributes.

[0045] When any common attribute includes a numerical attribute, the server will determine the numerical filtering interval according to the common filtering condition corresponding to the numerical attribute. For example, if the common filtering condition is that the age is under 35 years old, the numerical filtering interval is (0, 35). Then, the server will determine the regional value in the resume file through the regional semantics of the corresponding numerical attribute. For example, the determined regional value is 36. Then, the server will compare the regional value with the numerical filtering interval.

[0046] When the regional value is within the numerical screening interval, it indicates that the candidate's age meets the job requirements. At this time, the server will determine the attribute score corresponding to the general attribute as the first numerical score. When the regional value is outside the numerical screening interval, the server will further determine the maximum value and the minimum value in the numerical screening interval, thereby determining the maximum difference and the minimum difference between the regional value and the maximum value and the minimum value, respectively, and determine the difference with the smallest corresponding value as the evaluation difference.

[0047] For example, if the numerical screening interval is (0, 35), the maximum value in the numerical screening interval is 35 and the minimum value is 0. If the value in the area is 36, the maximum difference with the maximum value is 1, and the minimum difference with the minimum value is 36. At this time, the server will determine the evaluation difference as 1. Then, the server will multiply the evaluation difference with the retrieved difference coefficient to obtain the evaluation elimination value, and then perform the difference calculation between the first numerical score and the evaluation elimination value to obtain the second numerical score, and determine the second numerical score as the attribute score corresponding to the general attribute.

[0048] When any general attribute includes a noun attribute, the server will determine the noun filtering content based on the general filtering condition corresponding to the noun attribute. For example, if the general filtering condition is that the job type is software engineer, then the noun filtering content is software engineer. Then, the server will determine the regional noun in the resume file based on the regional semantics of the corresponding noun attribute, for example, the determined regional noun is software engineer or operation and maintenance engineer. Then, the server will compare the regional noun with the noun filtering content.

[0049] When the regional noun and the noun screening content have a matching relationship, it means that the candidate's job type meets the job requirements, so the server will determine the attribute score corresponding to the general attribute as the first noun score; when the regional noun and the noun screening content do not have a matching relationship, it means that the candidate's job type does not meet the job requirements, so the server will determine the attribute score corresponding to the general attribute as the second noun score.

[0050] In step S106, the following contents are included: The resume sequence is screened based on a preset screening ratio, and the resume files corresponding to the preset screening ratio are determined as a resume screening group. A resume template including filling areas corresponding to different unique attributes configured based on the management end is retrieved and sent to each talent end corresponding to each resume file in the resume screening group for filling.

[0051] For example, in this embodiment, the server will perform preliminary screening of the resume sequence based on a preset screening ratio, such as removing the resume files located in the last 20% of the resume sequence, thereby determining the resume files located in the first 80% of the resume sequence as a resume screening group.

[0052] The management end will configure a resume template based on the job recruitment situation. The resume template may contain some subjective questions, that is, it contains various filling areas corresponding to different unique attributes. After determining the resume screening group, the server will retrieve the resume template configured by the management end and send it to each talent end corresponding to each resume file in the resume screening group. When the talent end receives the resume template, it can fill in the corresponding content.

[0053] Furthermore, the above-mentioned "screening the resume sequence based on a preset screening ratio, determining the obtained resume files corresponding to the preset screening ratio as a resume screening group, retrieving a resume template including filling areas corresponding to different unique attributes configured based on the management end and sending it to each talent end corresponding to each resume file in the resume screening group for filling", further includes the following steps: Retrieving historical resume feedback data, wherein the historical resume feedback data includes different historical periods and the resume feedback ratios corresponding to the different historical periods; Calculate the average of each resume feedback ratio to obtain the mean ratio; Retrieving a first preset demand quantity, and calculating the ratio of the first preset demand quantity to the mean of the ratio to obtain a second preset demand quantity; Based on the second preset required number, the resume sequence is screened from high to low, and each resume file corresponding to the second preset required number is determined as a resume screening group; Retrieving a resume template configured based on the management end, wherein the resume template includes filling areas corresponding to different unique attributes; Based on the resume template, each resume file in the resume screening group is configured one by one, and each filling area in the resume template is updated based on each resume file; Send the resume templates that have been updated by the region to the corresponding talent terminals for filling out.

[0054] For example, in this embodiment, since it often happens that the talent end does not fill in and provide feedback accordingly after receiving the resume template, the number of feedbacks received by the management end about the resume template will be less than the expected number. In order to ensure that the number of feedbacks received by the management end can reach the expected number, the server will first retrieve the historical resume feedback data, which includes different historical periods and the feedback ratios of each resume corresponding to each different historical period. For example, during the recruitment period from January 2022 to February 2022, a total of 100 resume templates were issued, and a total of 80 resume templates were received for feedback. The resume feedback ratio for the historical period from January 2022 to February 2022 is 80%. Then, the server will average the feedback ratios of each resume to obtain the mean ratio.

[0055] Next, the server will retrieve the expected number of resume templates that the management end wants to receive feedback on, that is, the first preset requirement number, and calculate the ratio of the first preset requirement number to the proportion mean to obtain the second preset requirement number. In this way, after the server sends out a resume template with the second preset requirement number, the number of feedbacks obtained will definitely be greater than or equal to the first preset requirement number, which can ensure the number of resumes that can be screened by the subsequent interview end.

[0056] Then, the server will screen the resume sequence from high to low according to the second preset required number, and determine the resume files corresponding to the second preset required number as the resume screening group. Then, the resume template configured by the management end is retrieved, and the resume template includes various filling areas corresponding to different unique attributes.

[0057] Since some filling areas in the resume file of the talent end may match the filling areas of the resume template, in this case, the server will configure each resume file in the resume screening group one by one according to the resume template, and perform regional updates on each filling area in the resume template based on each resume file, and then send each resume template after regional updates to the corresponding talent end for filling. This embodiment can make it unnecessary to repeatedly fill in these filling areas in the resume template when some filling areas in the resume file of the talent end match the filling areas of the resume template.

[0058] Furthermore, the above-mentioned “updating each filling area located in the resume template based on each resume file” further includes the following steps: All remaining areas except the resume area corresponding to the general attribute in each resume file are respectively determined as the areas to be matched, and the semantic labels to be filled in corresponding to the filled-in areas are obtained; Performing semantic recognition on each to-be-matched region in each resume file to obtain the semantics of each to-be-matched region corresponding to each to-be-matched region; Match each to-be-matched region semantic with the to-be-filled semantic label one by one, and determine the to-be-matched region semantic that has a matching relationship with any to-be-filled semantic label as the matched region semantic; Based on the matching relationship, the to-be-matched area corresponding to the semantics of the matched area is mapped to the filling area corresponding to the semantic label to be filled, and all the mapped filling areas are converted from an initial to-be-filled state to a filled state; The area outlines of all filled-in areas in the resume template are configured with a first color, and the area outlines of all filled-in areas to be filled-in are configured with a second color different from the first color.

[0059] For example, in this embodiment, the server first determines all remaining areas in each resume file except the resume area corresponding to the general attribute as the to-be-matched areas, and obtains the to-be-filled semantic labels of each filled-in area, such as the obtained to-be-filled semantic labels are work experience, etc. Then, the server performs semantic recognition on each to-be-matched area in each resume file, thereby obtaining the to-be-matched area semantics of each to-be-matched area.

[0060] Next, the server will match the semantics of each area to be matched with the semantic labels to be filled in one by one. When the semantics of the area to be matched has a matching relationship with any filled-in content label, the server will determine the semantics of the area to be matched as the matched area semantics. Then, according to the matching relationship, the area to be matched corresponding to the semantics of the matched area will be mapped to the filled-in area corresponding to the semantic label to be filled in, thereby converting all the mapped filled-in areas from the initial state to be filled in to the filled-in state, avoiding repeated filling in by the subsequent talent end.

[0061] Then, the server will configure the area outlines of all filled-in areas in the resume template with a first color, and then configure the area outlines of all filled-in areas to be filled-in with a second color different from the first color, so that the talent side can more clearly understand which filling areas need to be filled in with corresponding content through color configuration.

[0062] In step S108, the following contents are included: In response to receiving a talent resume uploaded by any talent after filling in the resume template, the interview end is triggered to conduct a video interview with the talent end based on the filling areas corresponding to different unique attributes in the talent resume, and analyze and process the video interview based on a preset interview analysis strategy to obtain an interview score corresponding to the talent end.

[0063] For example, in this embodiment, when any talent terminal completes filling in the resume template, it will obtain the corresponding talent resume, so that the talent terminal can upload the talent resume. At this time, the server will receive the talent resume, thereby triggering the interview terminal to conduct a video interview with the talent terminal according to the filling areas corresponding to different unique attributes in the talent resume, such as "What is your personality like?", "What projects have you participated in?", then the interview terminal will ask the talent terminal corresponding questions in turn, and analyze and process this video interview based on the preset interview analysis strategy, so as to obtain the interview score corresponding to the talent terminal.

[0064] Further, the above-mentioned "in response to receiving a talent resume uploaded by any talent end after filling in the resume template, triggering the interview end to conduct a video interview with the talent end based on the filling areas corresponding to different unique attributes in the talent resume, and analyzing and processing the video interview based on a preset interview analysis strategy to obtain an interview score corresponding to the talent end" also includes the following steps: Create an interview record layer, and send the interview record layer to the interview end, wherein the interview record layer includes a resume display area, a video display area, and a key point record area; In response to the interview end establishing a video call with the talent end, mapping the video call to the video display area, and mapping the filling areas corresponding to different unique attributes in the talent resume to the resume display area for display; In response to the interaction of the interview end with any filling area, the state acquisition unit is triggered to collect the state of the video portion of the corresponding talent end located in the video display area, obtain the interview state corresponding to the talent end, and establish an active recording sub-area and a passive recording sub-area corresponding to the filling area in the key point recording area; Acquire the interview content fed back by the talent end based on the filling area, compare the interview content with the filling content in the filling area, and determine the subjective evaluation content corresponding to the filling area based on the comparison result; Adding the subjective evaluation content to the active recording sub-area, adding the interview status to the passive recording sub-area, and determining the active recording sub-area and the passive recording sub-area corresponding to the same filling area after the addition as a key content group; In response to completing the determination of the key content groups corresponding to each of the filled-in areas, performing content evaluation on each of the key content groups to obtain original scores corresponding to each of the key content groups; Obtain the scoring weights corresponding to the filled-in areas, and multiply the original scores by the corresponding scoring weights to obtain the scores for each key point; The scores of each key point are averaged, and the obtained average score is determined as the interview score corresponding to the talent end.

[0065] For example, in this embodiment, before the interview, the server will first create an interview record layer and send the interview record layer to the interview end, in which the interview record layer includes a resume display area, a video display area, and a key point record area. When the interview end establishes a video call with the talent end, the server will map the video call screen to the video display area, and map the filling areas corresponding to different unique attributes in the talent resume to the resume display area for display.

[0066] When the interviewer interacts with any filling area, the server triggers the status acquisition unit to collect the status of the video part of the corresponding talent end in the video display area, thereby obtaining the interview status of the talent end, and establishing active recording sub-areas and passive recording sub-areas corresponding to the filling area in the key point recording area in a top-down order, such as Figure 3As shown. The server will compare the interview content fed back by the talent with the content in the filling area, thereby generating subjective evaluation content corresponding to the filling area, and further adding it to the active record sub-area; then, the server will add the interview status of the talent to the passive record sub-area, and determine the active record sub-area and passive record sub-area corresponding to the same filling area after addition as the key content group.

[0067] After the key content groups corresponding to each filling area are determined, the server will evaluate the content of each key content group to obtain the original score of each key content group. Next, the server will obtain the score weights of each filling area, and multiply each original score by the corresponding score weight to obtain the score of each key point, and then average the score of each key point to obtain the average score. At this time, the server will determine the average score as the interview score corresponding to the talent end.

[0068] Furthermore, the above-mentioned “triggering the status acquisition unit to acquire the status of the video portion of the corresponding talent terminal located in the video display area to obtain the interview status corresponding to the talent terminal” further includes the following steps: The trigger state acquisition unit performs face recognition on the video portion to obtain a face recognition area corresponding to the talent end, and performs expression recognition and voice recognition based on the face recognition area to obtain an expression recognition state and a voice recognition state corresponding to the talent end; Retrieving a preset expression evaluation table, wherein the preset expression evaluation table includes expression evaluation scores corresponding to different expression states; Traversing the preset expression evaluation table to determine the expression evaluation score corresponding to the expression recognition state; Retrieving a preset sound evaluation table, wherein the preset sound evaluation table includes sound evaluation scores corresponding to different sound states; Traversing the preset sound evaluation table to determine the sound evaluation score corresponding to the sound recognition state; Based on the expression evaluation score and the voice evaluation score, a preset proportion is summed, and based on each summation result, an emotion score corresponding to each staff member is determined; Retrieving a status evaluation table, wherein the status evaluation table includes different score intervals and different interview states corresponding to the different score intervals; The state evaluation table is traversed to determine the interview state corresponding to the score interval containing the emotion score.

[0069] For example, in this embodiment, the server triggers the status acquisition unit to perform face recognition on the video part, thereby obtaining the face recognition area corresponding to the talent end, and performs expression recognition and voice recognition on the face recognition area respectively, thereby obtaining the expression recognition status and voice recognition status of the talent end.

[0070] Then, the server will retrieve a preset expression evaluation table, which includes expression evaluation scores corresponding to different expression states, and the server will determine the expression evaluation score corresponding to the expression recognition state in the preset expression evaluation table. Next, the server will retrieve a preset sound evaluation table, which includes sound evaluation scores corresponding to different sound states, and the server will determine the sound evaluation score corresponding to the sound recognition state in the preset sound evaluation table.

[0071] Finally, the server will sum the expression evaluation score and the voice evaluation score with a preset proportion to obtain the summed results, and determine the emotional score of each staff member based on the summed results. Then, the server will retrieve the status evaluation table, which includes different score intervals and different interview states corresponding to different score intervals. The server will determine the interview state corresponding to the score interval containing the emotional score in the status evaluation table.

[0072] Furthermore, in this embodiment, the above-mentioned "obtaining the interview content fed back by the talent end based on the filling area, and comparing the interview content with the filling content in the filling area, and determining the subjective evaluation content corresponding to the filling area based on the comparison result" may also include the following contents: Acquire the interview content fed back by the talent end based on the filling area, and perform semantic recognition on the interview content and the filling content respectively to obtain the interview semantics and the filling semantics corresponding to the interview content and the filling content respectively; The interview semantics is semantically compared with the filled-in semantics, and the subjective evaluation content corresponding to the filled-in area is determined based on the comparison result, wherein the subjective evaluation content includes positive evaluation and negative evaluation.

[0073] For example, in this embodiment, when the interview end conducts an interview with the talent end based on the talent resume, the interview end will ask the talent end corresponding interview questions based on the filling area in the talent resume one by one, and after obtaining the interview content fed back by the talent end, the server will perform corresponding semantic recognition on the interview content and the filling content in the filling area, so as to obtain the corresponding interview semantics and filling semantics; finally, the server can perform semantic comparison on the interview semantics and the filling semantics, so as to determine whether the interview content answered by the talent end matches the filling content in the filling area, that is, it can be judged whether the filling content filled in by the talent end for the filling area is correct, so as to determine the integrity of the resume corresponding to the talent end, thereby improving the interview efficiency of the interview process in an automated manner.

[0074] Furthermore, in this embodiment, the above-mentioned “in response to completing the determination of the key content groups corresponding to each fill-in area, performing content evaluation on each key content group, and obtaining original scores corresponding to each key content group” may also include the following method steps: In response to completing the determination of the key content groups corresponding to each filling area, obtaining the subjective evaluation content and the interview status respectively included in each key content group; In response to the subjective evaluation content included in any key content group being a negative evaluation, determining the retrieved first preset score as an original score corresponding to the key content group; In response to the subjective evaluation content included in any key content group being a positive evaluation, determining the retrieved second preset score as the subjective score corresponding to the subjective evaluation content, and retrieving a preset state table, wherein the preset state table includes different interview states and state scores corresponding to the different interview states; Obtaining an interview state that is in the same key content group as the subjective evaluation content, and traversing the preset state table to determine a state score corresponding to the interview state; The second preset score and the state score are summed up and the sum result is determined as the original score corresponding to the key content group.

[0075] For example, in this embodiment, after the server has completed determining the key content groups corresponding to each filling area, the server can obtain the subjective evaluation content and the interview status respectively included in each key content group, thereby further determining the original score corresponding to each key content group respectively according to the subjective evaluation content and the interview status; Among them, it can be explained that, since the subjective evaluation content is determined based on the semantic comparison between the interview semantics and the filling semantics, therefore, when the interview semantics is different from the filling semantics, it indicates that the interview content fed back by the talent end during the interview is different from the filling content filled in on the resume. At this time, it can be judged that the filling or interview based on the filling content by the talent end may contain deception. Therefore, the subjective evaluation content can be determined as a negative evaluation, and the first preset score retrieved is further determined as the original score corresponding to the key content group; and when the interview semantics is the same as the filling semantics, it indicates that the interview content fed back by the talent end during the interview is the same as the filling content filled in on the resume. At this time, it can be judged that the filling or interview based on the filling content by the talent end may not contain deception. Therefore, the subjective evaluation content can be determined as a positive evaluation, and the second preset score is retrieved to determine it as the subjective score corresponding to the subjective evaluation content; Here, in the actual implementation process, according to psychology, anyone who lies may have different facial expressions or movements to cover up the uneasiness caused by lying, such as touching the head, touching the nose, squinting the eyes, etc.; based on this feature, it is possible to further determine whether the talent side has a deceptive component by combining the obtained interview status of the talent side, that is, calling the corresponding preset status table, wherein the preset status table includes various interview states and status scores corresponding to various interview states. Here, the status scores corresponding to lying facial expressions and movements are generally lower, while the status scores corresponding to non-lying facial expressions and movements are generally higher.

[0076] By traversing the preset status table accordingly with the interview status of the corresponding talent end, the status score corresponding to the interview status can be determined, so that the subjective score of the corresponding subjective evaluation content and the status score can be further summed up and the sum result can be determined as the original score corresponding to the key content group.

[0077] It can be explained that, in this embodiment, when the corresponding subjective evaluation content is a positive evaluation, the corresponding original score can be used to roughly determine whether the talent side may have deceptive elements, that is, a corresponding preset evaluation threshold can be set. When the original score is greater than or equal to the preset evaluation threshold, it can be roughly determined that the talent side has no deceptive elements, and when the original score is less than the preset evaluation threshold, it can be roughly determined that the talent side is likely to have deceptive elements, thereby reminding the interview side to conduct further content inspection based on the filled content of the filling area to improve the interview efficiency and interview accuracy.

[0078] According to the solution of the present invention, the server will first extract the various different areas such as the personal photo area in each resume file sent by each different talent terminal, so as to obtain the resume areas corresponding to different general attributes configured by the management terminal, so as to facilitate the subsequent management terminal to screen each resume file. The server will also evaluate the resume area corresponding to the general attribute, and then obtain the resume screening score corresponding to the different resume files according to each resume evaluation value. Then, the server will rank each resume file in the order of high to low resume screening scores to obtain a resume sequence, so that the subsequent management terminal can check which resumes are high-quality resumes according to the resume sequence. Then, the server will perform a preliminary screening of the resume sequence based on the preset screening ratio, so as to determine the obtained resume files corresponding to the preset screening ratio as a resume screening group. And the management terminal can configure a resume template containing each filling area corresponding to different unique attributes according to the job recruitment situation. The server will retrieve the resume template and send it to each talent terminal corresponding to each resume file in the resume screening group for filling. When any talent end completes filling in the resume template, the corresponding talent resume will be obtained. When the server receives the talent resume uploaded by the talent end, it will trigger the interview end to conduct a video interview with the talent end according to the filling areas corresponding to different unique attributes in the talent resume, and analyze and process this video interview based on a preset interview analysis strategy, so as to obtain an interview score corresponding to the talent end. The present invention can improve the efficiency of resume screening, and obtain an interview score for the video interview, which can provide a certain reference to the management end.

[0079] Another embodiment of the present invention provides a talent resume screening and matching device based on intelligence, Figure 4 The corresponding device block diagram includes: The region extraction module is configured to perform region extraction on each resume file sent by each different talent terminal, and obtain the resume regions corresponding to different general attributes configured by the management terminal; A sequence generation module is configured to evaluate resumes based on resume areas corresponding to different general attributes in each resume file, and to rank the resume files from high to low based on the obtained resume screening scores to obtain a resume sequence; A resume screening module is configured to screen the resume sequence based on a preset screening ratio, determine the obtained resume files corresponding to the preset screening ratio as a resume screening group, retrieve a resume template including filling areas corresponding to different unique attributes configured based on the management end, and send it to each talent end corresponding to each resume file in the resume screening group for filling; The interview scoring module is configured to respond to the receipt of a talent resume uploaded by any talent end after filling in the resume template, trigger the interview end to conduct a video interview with the talent end based on the filling areas corresponding to different unique attributes in the talent resume, and analyze and process the video interview based on a preset interview analysis strategy to obtain an interview score corresponding to the talent end.

[0080] In the specification provided herein, algorithms and displays are not inherently related to any particular computer, virtual system or other device. Various general systems can also be used together with the examples of the present invention. From the above description, it is obvious that the structure required for constructing such systems. In addition, the present invention is not directed to any specific programming noun. It should be understood that various programming nouns can be used to implement the content of the present invention described herein, and the description of specific nouns above is to disclose the preferred embodiment of the present invention.

[0081] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0082] Similarly, it should be understood that in order to streamline the present disclosure and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof.

[0083] Those skilled in the art will appreciate that the modules or units or components of the devices in the examples disclosed herein may be arranged in the devices described in the embodiment, or alternatively may be located in one or more devices different from the devices in the examples. The modules in the foregoing examples may be combined into one module or may be divided into multiple submodules.

[0084] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and furthermore may be divided into a plurality of submodules or subunits or subcomponents.

[0085] Furthermore, those skilled in the art will appreciate that although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features from different embodiments is meant to be within the scope of the present invention and to form different embodiments.

[0086] In addition, some of the embodiments are described herein as methods or combinations of method elements that can be implemented by a processor of a computer system or by other devices that perform the functions. Therefore, a processor with necessary instructions for implementing the method or method elements forms a device for implementing the method or method elements. In addition, the elements described herein of the device embodiments are examples of devices for implementing the functions performed by the elements for the purpose of implementing the invention.

[0087] As used herein, unless otherwise specified, the use of ordinal numbers "first," "second," "third," etc. to describe common objects merely indicates that different instances of similar objects are involved, and is not intended to imply that the objects so described must have a given order in time, space, order, or in any other manner.

[0088] Although the present invention has been described according to a limited number of embodiments, it will be apparent to those skilled in the art, with the benefit of the above description, that other embodiments are contemplated within the scope of the invention thus described. In addition, it should be noted that the nouns used in this specification are primarily selected for readability and teaching purposes, rather than for explaining or limiting the subject matter of the present invention.

Claims

1. A talent resume screening and matching method based on intelligence, characterized in that: The following steps are involved: Extract regions from the resume files sent by different talent terminals to obtain resume regions corresponding to different general attributes based on the configuration of the management terminal; Resume evaluation is performed based on resume areas corresponding to different general attributes respectively located in each resume file, and resume ranking of each resume file is performed from high to low based on the obtained resume screening scores to obtain a resume sequence; Screening the resume sequence based on a preset screening ratio, determining the obtained resume files corresponding to the preset screening ratio as a resume screening group, retrieving a resume template including filling areas corresponding to different unique attributes configured based on the management end, and sending the template to each talent end corresponding to each resume file in the resume screening group for filling; In response to receiving a talent resume uploaded by any talent after filling in the resume template, the interview end is triggered to conduct a video interview with the talent end based on the filling areas corresponding to different unique attributes in the talent resume, and analyze and process the video interview based on a preset interview analysis strategy to obtain an interview score corresponding to the talent end.

2. The method for screening and matching talent resumes based on intelligence according to claim 1 is characterized in that: Perform regional extraction on each resume file sent by different talent terminals to obtain resume regions corresponding to different general attributes based on the configuration of the management terminal, including: The structures of the resume files sent by different talent terminals are obtained respectively, and the file segments in the same resume file are obtained, and the resume regions formed by the adjacent file segments are determined as region extraction groups; Perform pixel identification on each resume area in the area extraction group, and remove all resume areas that do not contain any pixel from the area extraction group to obtain an updated area extraction group; Performing character recognition on each resume region in the updated region extraction group, and obtaining common key contents corresponding to different common attributes configured by the management end; Each resume area is traversed, and a resume area including any of the universal key contents is determined as a universal attribute corresponding to the universal key content, thereby obtaining resume areas corresponding to different universal attributes.

3. The method for screening and matching talent resumes based on intelligence according to claim 2 is characterized in that: Each resume area formed by each adjacent file line segment is determined as an area extraction group, including: Performing coordinate processing on each resume file respectively, and determining each area coordinate group corresponding to each resume area based on each resume file after coordinate processing; The regional coordinate groups corresponding to the coordinate extreme values ​​in each regional coordinate group are removed, and all the remaining regional coordinate groups are compared in pairs, and it is determined whether there is an enclosing relationship between the regional coordinate groups based on the comparison results; The region coordinate groups having an enclosing relationship are merged into a resume region, the region coordinate groups not having an enclosing relationship are respectively determined as different resume regions, and the obtained resume regions are determined as region extraction groups.

4. The method for screening and matching talent resumes based on intelligence according to claim 1 is characterized in that: Resume evaluation is performed based on resume areas corresponding to different general attributes in each resume file, and resume ranking of each resume file is performed from high to low based on the obtained resume screening scores to obtain a resume sequence, including: Obtaining general screening conditions corresponding to different general attributes configured on the management side, and obtaining general area semantics corresponding to the resume areas corresponding to different general attributes; Based on any general screening condition, a regional evaluation is performed on the general regional semantics corresponding to the same general attribute, and based on the evaluation result, a score of each attribute corresponding to each general attribute is determined; Based on the average calculation of each attribute score, each resume screening score corresponding to each resume file is obtained, and based on the obtained resume screening score, each resume file is ranked from high to low to obtain a resume sequence, wherein the resume sequence includes each resume file and a ranking number corresponding to each resume file.

5. The method for screening and matching talent resumes based on intelligence according to claim 4 is characterized in that: Based on any general screening condition, a regional evaluation is performed on the general regional semantics corresponding to the same general attribute, and based on the evaluation result, the attribute scores corresponding to each general attribute are determined, including: When any of the common attributes include numeric attributes, determining a numerical filtering interval based on a general filtering condition corresponding to the numerical attribute, determining a region value based on a general region semantics corresponding to the numerical attribute, and comparing the region value with the numerical filtering interval; When the region value is within the value screening interval, determining the attribute score corresponding to the general attribute as the first value score; When the region value is outside the value screening interval, a maximum value and a minimum value in the value screening interval are determined, and a maximum difference and a minimum difference between the region value and the maximum value and the minimum value are determined respectively; The difference with the smallest corresponding value is determined as the evaluation difference, and the evaluation difference is multiplied by the retrieved difference coefficient to obtain the evaluation elimination value; Calculating the difference between the first numerical score and the evaluation elimination value, and determining the obtained second numerical score as the attribute score corresponding to the general attribute; and / or When any common attribute includes a noun attribute, Determining noun screening content based on the general screening conditions corresponding to the noun attributes, determining regional nouns based on the general regional semantics corresponding to the noun attributes, and comparing the regional nouns with the noun screening content; When the regional noun has a matching relationship with the noun screening content, determining the attribute score corresponding to the general attribute as the first noun score; When the regional noun does not have a matching relationship with the noun screening content, the attribute score corresponding to the general attribute is determined as the second noun score.

6. The method for screening and matching talent resumes based on intelligence according to claim 1 is characterized in that: The resume sequence is screened based on a preset screening ratio, and each resume file corresponding to the preset screening ratio is determined as a resume screening group. A resume template including filling areas corresponding to different unique attributes configured based on the management end is retrieved and sent to each talent end corresponding to each resume file in the resume screening group for filling, including: Retrieving historical resume feedback data, wherein the historical resume feedback data includes different historical periods and the resume feedback ratios corresponding to the different historical periods; Calculate the average of each resume feedback ratio to obtain the mean ratio; Retrieving a first preset demand quantity, and calculating the ratio of the first preset demand quantity to the mean of the ratio to obtain a second preset demand quantity; Based on the second preset required number, the resume sequence is screened from high to low, and each resume file corresponding to the second preset required number is determined as a resume screening group; Retrieving a resume template configured based on the management end, wherein the resume template includes filling areas corresponding to different unique attributes; Based on the resume template, each resume file in the resume screening group is configured one by one, and each filling area in the resume template is updated based on each resume file; Send the resume templates that have been updated by the region to the corresponding talent terminals for filling out.

7. The method for screening and matching talent resumes based on intelligence according to claim 6 is characterized in that: Based on each resume file, each filling area of ​​the resume template is updated, including: Determine all remaining areas in each resume file except the resume area corresponding to the general attribute as the areas to be matched, and obtain the semantic labels to be filled in corresponding to the areas to be filled in; Performing semantic recognition on each to-be-matched region in each resume file to obtain the semantics of each to-be-matched region corresponding to each to-be-matched region; Match each to-be-matched region semantic with the to-be-filled semantic label one by one, and determine the to-be-matched region semantic that has a matching relationship with any to-be-filled semantic label as the matched region semantic; Based on the matching relationship, the to-be-matched area corresponding to the semantics of the matched area is mapped to the filling area corresponding to the semantic label to be filled, and all the mapped filling areas are converted from an initial to-be-filled state to a filled state; The area outlines of all filled-in areas in the resume template are configured with a first color, and the area outlines of all filled-in areas to be filled-in are configured with a second color different from the first color.

8. The method for screening and matching talent resumes based on intelligence according to claim 1 is characterized in that: In response to receiving a resume uploaded by any talent terminal after filling in the resume template, triggering the interview terminal to conduct a video interview with the talent terminal based on the filling areas corresponding to different unique attributes in the talent resume, and analyzing and processing the video interview based on a preset interview analysis strategy to obtain an interview score corresponding to the talent terminal, including: Create an interview record layer, and send the interview record layer to the interview end, wherein the interview record layer includes a resume display area, a video display area, and a key point record area; In response to the interview end establishing a video call with the talent end, mapping the video call to the video display area, and mapping the filling areas corresponding to different unique attributes in the talent resume to the resume display area for display; In response to the interaction of the interview end with any filling area, the state acquisition unit is triggered to collect the state of the video portion of the corresponding talent end located in the video display area, obtain the interview state corresponding to the talent end, and establish an active recording sub-area and a passive recording sub-area corresponding to the filling area in the key point recording area; Acquire the interview content fed back by the talent end based on the filling area, compare the interview content with the filling content in the filling area, and determine the subjective evaluation content corresponding to the filling area based on the comparison result; Adding the subjective evaluation content to the active recording sub-area, adding the interview status to the passive recording sub-area, and determining the active recording sub-area and the passive recording sub-area corresponding to the same filling area after the addition as a key content group; In response to completing the determination of the key content groups corresponding to each of the filled-in areas, performing content evaluation on each of the key content groups to obtain original scores corresponding to each of the key content groups; Obtain the scoring weights corresponding to the filled-in areas, and multiply the original scores by the corresponding scoring weights to obtain the scores for each key point; The scores of each key point are averaged, and the obtained average score is determined as the interview score corresponding to the talent end.

9. The method for screening and matching talent resumes based on intelligence according to claim 8 is characterized in that: The triggering state acquisition unit acquires the state of the video portion of the corresponding talent terminal located in the video display area to obtain the interview state corresponding to the talent terminal, including: The trigger state acquisition unit performs face recognition on the video portion to obtain a face recognition area corresponding to the talent end, and performs expression recognition and voice recognition based on the face recognition area to obtain an expression recognition state and a voice recognition state corresponding to the talent end; Retrieving a preset expression evaluation table, wherein the preset expression evaluation table includes expression evaluation scores corresponding to different expression states; Traversing the preset expression evaluation table to determine the expression evaluation score corresponding to the expression recognition state; Retrieving a preset sound evaluation table, wherein the preset sound evaluation table includes sound evaluation scores corresponding to different sound states; Traversing the preset sound evaluation table to determine the sound evaluation score corresponding to the sound recognition state; Based on the expression evaluation score and the voice evaluation score, a preset proportion is summed, and based on each summation result, an emotion score corresponding to each staff member is determined; Retrieving a status evaluation table, wherein the status evaluation table includes different score intervals and different interview states corresponding to the different score intervals; The state evaluation table is traversed to determine the interview state corresponding to the score interval containing the emotion score.

10. An intelligent talent resume screening and matching device, characterized in that: include: The region extraction module is configured to perform region extraction on each resume file sent by each different talent terminal, and obtain the resume regions corresponding to different general attributes configured by the management terminal; A sequence generation module is configured to evaluate resumes based on resume areas corresponding to different general attributes in each resume file, and to rank the resume files from high to low based on the obtained resume screening scores to obtain a resume sequence; A resume screening module is configured to screen the resume sequence based on a preset screening ratio, determine the obtained resume files corresponding to the preset screening ratio as a resume screening group, retrieve a resume template including filling areas corresponding to different unique attributes configured based on the management end, and send it to each talent end corresponding to each resume file in the resume screening group for filling; The interview scoring module is configured to respond to the receipt of a talent resume uploaded by any talent end after filling in the resume template, trigger the interview end to conduct a video interview with the talent end based on the filling areas corresponding to different unique attributes in the talent resume, and analyze and process the video interview based on a preset interview analysis strategy to obtain an interview score corresponding to the talent end.

Images