Material work order data processing method, device, equipment and storage medium

Abstract

The present invention relates to the field of data processing technology, and discloses a material work order data processing method, device, equipment, and storage medium. The method includes: obtaining a demand file and reference material corresponding to the material work order; performing text recognition and demand extraction on the demand file to obtain a material demand result, and performing material style recognition on the reference material to obtain a style recognition result; obtaining a final demand result; based on a distribution rule, distributing the final demand result to a task queue corresponding to a matched material designer; receiving a material document returned by the task queue; performing transmission format conversion processing on the material document, and pushing the converted material document to the demander corresponding to the material work order. Therefore, the present invention realizes automatic identification of the final demand result, automatic distribution of tasks, and automatic conversion to the demander, thereby improving customer satisfaction with the pushed material documents and ensuring high matching of the designed material documents.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a material work order data processing method, device, equipment and storage medium. Background Art

[0002] At present, with the development of media technology, we can see advertisements everywhere, and advertisements are distributed in various forms. Most advertisements are based on images and / or videos as materials. Different customers have different needs for material delivery strategies. As the number of advertisements increases, the demand for materials increases, and the strategies for material delivery also increase. In existing technical solutions, the demand for materials and the strategies for material delivery are mostly communicated between advertising optimizers and customers. Advertising optimizers manually interpret the demand for materials and the strategies for material delivery, and then convey them to advertising designers for design. Since different advertising optimizers interpret the content in various ways, and an advertising optimizer connects with many customers, it is easy to interfere with each other, resulting in material confusion, and the influence of subjective factors such as incorrect interpretation and missing information, resulting in chaos in the management of material development and material delivery strategies for each advertisement, low accuracy and correctness, resulting in the produced materials and material delivery strategies not fully meeting customer needs, resulting in low customer satisfaction. Summary of the Invention

[0003] The present invention provides a material work order data processing method, device, computer equipment and storage medium, which can automatically identify the final demand results, automatically distribute them to the task queue corresponding to the matching material designer, and automatically convert the received material documents, thereby greatly reducing the manual workload and improving the customer satisfaction of the pushed material documents.

[0004] A material work order data processing method, comprising:

[0005] Obtain the requirement documents and reference materials corresponding to the material work order;

[0006] Performing text recognition and demand extraction on the demand document to obtain a material demand result, and performing material style recognition on the reference material to obtain a style recognition result;

[0007] Recording the material demand result and the style recognition result as the final demand result;

[0008] Based on the distribution rules, the final demand result is distributed to the task queue corresponding to the material designer matching the final demand result;

[0009] Receive the material document returned by the task queue; the material document is a document created by the material designer according to the final demand result;

[0010] The material document is converted into a transmission format corresponding to the material work order, and the converted material document is pushed to the demander corresponding to the material work order.

[0011] A material work order data processing device, comprising:

[0012] The acquisition module is used to obtain the requirement files and reference materials corresponding to the material work order;

[0013] an identification module for performing text recognition and demand extraction on the demand document to obtain a material demand result, and performing material style recognition on the reference material to obtain a style recognition result;

[0014] A recording module, configured to record the material demand result and the style recognition result as a final demand result;

[0015] A distribution module, configured to distribute the final demand result to a task queue corresponding to a material designer matching the final demand result based on a distribution rule;

[0016] A receiving module, configured to receive the material document returned by the task queue; the material document is a document created by the material designer according to the final demand result;

[0017] The conversion module is used to convert the material document into a transmission format corresponding to the material work order, and push the converted material document to the demander corresponding to the material work order.

[0018] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the material work order data processing method are implemented.

[0019] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned material work order data processing method.

[0020] The material work order data processing method, device, computer equipment and storage medium provided by the present invention obtain the demand file and reference material corresponding to the material work order; perform text recognition and demand extraction on the demand file to obtain a material demand result, and perform material style recognition on the reference material to obtain a style recognition result; record the material demand result and the style recognition result as the final demand result; based on the distribution rule, distribute the final demand result to the task queue corresponding to the material designer matching the final demand result; receive the material document returned by the task queue; the material document is a document created by the material designer according to the final demand result; convert the material document into a transmission format corresponding to the material work order, and push the converted material document to the demander corresponding to the material work order. Therefore, the method realizes the automatic identification of the final demand result by using text recognition, demand extraction and material style recognition, and automatically distributes it to the task queue corresponding to the matching material designer, and automatically converts the received material document to send it to the demander. Therefore, the manual workload is greatly reduced, the customer satisfaction of the pushed material documents is improved, and the high matching of the designed material documents is guaranteed. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0022] Figure 1 This is a schematic diagram of an application environment of a material work order data processing method according to an embodiment of the present invention;

[0023] Figure 2 This is a flow chart of a material work order data processing method according to one embodiment of the present invention;

[0024] Figure 3 This is a flowchart of step S20 of the material work order data processing method in one embodiment of the present invention;

[0025] Figure 4 This is a principle block diagram of a material work order data processing device according to one embodiment of the present invention;

[0026] Figure 5 This is a principle block diagram of an identification module of a material work order data processing device in one embodiment of the present invention;

[0027] Figure 6 is a schematic diagram of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0029] The material work order data processing method provided by the present invention can be applied in Figure 1 In an application environment, a client (computer device or terminal) communicates with a server via a network. The client (computer device or terminal) includes, but is not limited to, various personal computers, laptops, smartphones, tablet computers, and portable wearable devices. The server can be implemented as a standalone server or a server cluster consisting of multiple servers.

[0030] In one embodiment, if Figure 2 As shown, a material work order data processing method is provided, and its technical solution mainly includes the following steps S10-S60:

[0031] S10, obtaining the requirement file and reference materials corresponding to the material work order.

[0032] It is understandable that when creating an advertisement, a series of materials corresponding to the advertisement need to be created. The materials are media files that can be displayed, such as images and / or videos, which are consistent with the demand file and reference materials of the advertisement. The material work order is a unique work order number assigned to a series of materials under the advertisement. One material work order corresponds to the demand file of one advertisement and one reference material. The demand file is a related file involving the demand proposed by the advertisement. The demand file is a file that can display the demand, such as images, text, and videos. The format of the demand file can be determined according to the method of collecting the demand. If the demand is collected by recording, the format of the demand file is an audio file; if the demand is collected by text recording, the format of the demand file is a text file; if the demand is collected by shooting a video, the format of the demand file is a video file.

[0033] The reference material is a material of a style to be referenced, and the reference material provides a method for generating a material of a similar style that meets the requirements.

[0034] S20, performing text recognition and demand extraction on the demand document to obtain a material demand result, and performing material style recognition on the reference material to obtain a style recognition result.

[0035] It can be understood that the text recognition is the process of identifying the text content of the input file, and the demand extraction is the process of extracting keywords related to the demand from the text content obtained after text recognition. The process of performing text recognition and demand extraction on the demand file includes: first, performing text conversion of the demand file into a corresponding format and converting it into text content in a text format; then, using the trained material keyword detection model to extract keywords related to the demand elements from the text content, wherein the network structure of the material keyword detection model can be a network structure based on Bi-LSTM (also known as: Bi-directional Long Short-Term Memory, bidirectional long short-term memory algorithm), and the bidirectional long short-term memory algorithm is an algorithm that combines the information of the input sequence in both forward and backward directions on the basis of LSTM (Long short-term memory); finally, the identified keywords are classified by the keyword detection model, and the categories of different dimensions are aggregated to obtain a set of keywords of each dimension, thereby obtaining the material demand result.

[0036] Understandably, the material style recognition is a process of extracting style features of the material and clustering the extracted style features to identify the style of the material. The process of material style recognition of the reference material can be achieved through a trained image style detection model. The training process of the image style detection model is to identify samples marked with style labels collected historically to obtain recognition results, and then iteratively update the parameters of the image style detection model according to the loss value between the style label and the recognition result until the loss value reaches the convergence condition and the training is stopped. The network structure of the image style detection model can be set according to demand. For example, the network structure of the image style detection model is a CNN network structure or a VGG16 network structure. The process of material style recognition of the reference material can be: first, the reference material is preprocessed by the image style detection model. The texture features of the lines in the reference material can be enhanced by image preprocessing. The method is a process of processing the reference material using an image enhancement algorithm, wherein the image enhancement algorithm includes image denoising, increasing clarity (contrast), grayscale or obtaining image edge features or performing convolution, binarization, histogram equalization, Laplace transform and gamma transform on the image, thereby obtaining a preprocessed image; secondly, style feature extraction is performed on the preprocessed image, and the style feature extraction can be a parametric texture feature extraction method based on statistical distribution or a non-parametric texture feature extraction method based on Markov random field, wherein the parametric texture feature extraction method based on statistical distribution is a statistical method that parameterizes the extracted texture in a statistical way, and the non-parametric texture feature extraction method based on Markov random field is to extract texture features using a Markov network; finally, the extracted texture features are classified, and the probability distribution of the categories of each material style is predicted, and the categories of the material styles that exceed the preset threshold are summarized to obtain a style recognition result.

[0037] Among them, the LSTM algorithm forgets the information in the cell state and remembers new information so that information useful for subsequent calculations can be transmitted, while useless information is discarded, and the hidden layer state is output at each time step. The forgetting, memory and output are obtained by comprehensive control of the forget gate, memory gate and output gate calculated based on the hidden layer state of the previous moment and the current input. The style features are features related to the style of the texture presented by the lines or regions in the image.

[0038] In one embodiment, if Figure 3 As shown, in step S20, that is, performing text recognition and demand extraction on the demand document to obtain material demand results, the following steps are included:

[0039] S201: Perform text conversion on the requirement document to obtain a requirement text.

[0040] It can be understood that the text conversion process is a process of identifying the file format of the input requirement file, and then converting the requirement file through a text conversion algorithm corresponding to the file format, and recording the converted text content as the requirement text, which reflects the text content contained in the requirement file.

[0041] In one embodiment, step S201, i.e., converting the requirement document into a text file to obtain the requirement text, includes:

[0042] Perform file format recognition on the demand file to identify the file format of the demand file.

[0043] It can be understood that the file format recognition is a process of identifying the format type of the input demand file. The process of file format recognition is to obtain the extension of the demand file and distinguish the file format according to the obtained extension, that is, different extensions correspond to different file formats, and one file format can correspond to multiple extensions. The file formats include text format, audio format, video format, image format, etc. For example, the extension of jpg corresponds to the image format, and the extension of txt corresponds to the text format, thereby identifying the file format of the demand file. In one embodiment, when it is detected that the file format is a text format, the demand file is opened, and the text content in the demand file is extracted, and the extracted text content is recorded as the demand text.

[0044] The requirement file is converted into a text format corresponding to the file format to obtain the requirement text.

[0045] It can be understood that different file formats correspond to different text conversion algorithms, that is, audio format corresponds to the algorithm in speech recognition technology, video format corresponds to the algorithm in audio extraction technology and the algorithm in speech recognition technology, and image format corresponds to the algorithm in optical character recognition technology (OCR technology). When it is detected that the file format is an audio format, the algorithm in speech recognition technology is used to perform speech-to-text conversion on the required file to obtain the required text; when it is detected that the file format is a video format, the algorithm in audio extraction technology is used to extract the audio stream of the required file, and speech recognition technology is used to identify the text content in the audio stream to obtain the required text; when it is detected that the file format is an image format, the algorithm in OCR technology is used to perform text extraction on the required file to obtain the required text.

[0046] Among them, the speech recognition technology (Automatic Speech Recognition, ASR) is a technology that converts human speech into text, that is, extracting voiceprint features from an input audio format file, that is, extracting voiceprint features with Mel-frequency cepstral coefficients (MFCC) in the file, and identifying the pronounced words corresponding to the voiceprint features through the voiceprint features, so as to convert the corresponding text content. The audio extraction technology is a technology that plays an input video format file through an audio extractor, collects the content of the audio stream during the playback process, and filters the noise of the collected audio stream content. The optical character recognition technology, also known as Optical Character Recognition (OCR) technology, uses an optical method to photograph the text copied in a paper document into a black and white dot matrix image file, and converts the text in the image into text through a trained text recognition network.

[0047] The present invention realizes the recognition of the file format of the demand file by performing file format recognition on the demand file; performing text conversion corresponding to the file format on the demand file to obtain the demand text. In this way, the file format of the demand file can be automatically recognized, and the corresponding text conversion can be performed to obtain the demand text without manual conversion into text, thereby meeting the requirements for the diversity of file formats of the demand file, greatly reducing the workload of manual conversion, and improving the efficiency of demand file conversion, thereby ensuring the correctness and accuracy of the conversion.

[0048] In one embodiment, converting the requirement document into a text format corresponding to the file format to obtain the requirement text includes:

[0049] When it is detected that the file format is an audio format, speech recognition technology is used to perform speech-to-text conversion on the required file to obtain the required text.

[0050] Understandably, the Automatic Speech Recognition (ASR) technology is a technology that converts human speech into text, that is, extracts voiceprint features from an input audio format file, that is, extracts voiceprint features with Mel-frequency cepstral coefficients (MFCC) in the file, and uses the voiceprint features to identify the pronounced words corresponding to the voiceprint features, thereby converting the corresponding text content. The speech-to-text conversion process is a process of conversion using the algorithm in the speech recognition technology.

[0051] When it is detected that the file format is a video format, the audio extraction technology and the speech recognition technology are used to perform video conversion on the demand file to obtain the demand text.

[0052] It can be understood that the audio extraction technology is a technology that plays the input video format file through an audio extractor, collects the content of the audio stream during the playback process, and performs noise filtering on the collected audio stream content. The video conversion process is to first use the audio extraction technology to extract the audio stream content in the required file, and then use the speech recognition technology to convert the audio stream content into text, and finally obtain the required text.

[0053] When it is detected that the file format is an image format, the OCR technology is used to extract text from the demand file to obtain the demand text.

[0054] It can be understood that the optical character recognition technology, also known as Optical Character Recognition (OCR) technology, uses an optical method to capture the text copied in a paper document into a black and white dot image file, and converts the text in the image into text through a trained text recognition network. The text extraction process is the process of using optical character recognition technology to identify the text in the required document.

[0055] In this way, by detecting demand files of different formats, the corresponding technology can be automatically selected and converted to obtain the demand text without the need for manual identification and conversion one by one, thereby improving the efficiency and accuracy of demand text conversion.

[0056] S202: performing material requirement identification on the requirement text to obtain a requirement identification result, and performing material strategy identification on the requirement text to obtain a strategy identification result.

[0057] Understandably, the material demand identification can be identified by the material factor extraction sub-network in the material keyword detection model, the material factor extraction sub-network is a neural network based on the Bi-LSTM network structure, the material factor extraction sub-network is a network trained by historical sentences containing text related to various factors, that is, by inputting sentence samples and material factor keywords associated with the sentence samples into the sub-network, the keywords in the sentence samples are identified by the Bi-LSTM semantic recognition method, the keywords are compared with the corresponding material factor keywords, the difference between the two is calculated, and the parameters of the sub-network are iteratively updated according to the difference, and the material factor extraction sub-network is continuously trained to complete the material factor extraction sub-network, the material factor extraction sub-network can be a multi-task learning neural network, that is, the material factor extraction sub-network The network includes multiple branch networks, and each branch network learns the category or feature extraction of a material factor through deep learning. The feature extraction of the material factor includes platform entity features, location entity features, time features, delivery mode entity features, specification features, layout features, duration features and other features related to the material factors. The features extracted by each branch network are classified by their respective fully connected layers to obtain the category results output by each branch network. Finally, the category results output by all the branch networks are summarized to obtain the demand identification result. The demand identification result includes the category of the advertising delivery platform, the location of the advertising delivery, the time of the advertising delivery, the mode of the advertising delivery, the material size, the material layout, the material duration and other category results related to the demand for advertising delivery or materials.

[0058] Among them, the material strategy identification can be identified by the material strategy identification sub-network in the material keyword detection model. The material strategy identification sub-network is a neural network based on the Bi-LSTM network structure. The material strategy identification sub-network is based on the historical sentences containing the category of advertising delivery platforms related to the material strategy, the location of advertising delivery, the time of advertising delivery, the mode of advertising delivery, etc., and by learning the semantic correlation between the category of advertising delivery platforms, the location of advertising delivery, the time of advertising delivery, and the mode of advertising delivery in the sentence, the probability distribution of the combination between the category of advertising delivery platforms, the location of advertising delivery, the time of advertising delivery, and the mode of advertising delivery is predicted, thereby obtaining the combination with the highest probability, and comparing the combination with the real advertising delivery in the sentence. The material strategy identification sub-network is trained to compare the combination of the category of the advertising delivery platform, the location of the advertising delivery, the time of the advertising delivery, and the mode of the advertising delivery, and the convergence is continuously completed until the convergence is completed to obtain the trained material strategy identification sub-network. The process of the material strategy identification is to predict the probability of various combinations of the category of the advertising delivery platform, the location of the advertising delivery, the time of the advertising delivery, and the mode of the advertising delivery for the input demand text, so as to obtain the combination corresponding to the highest probability, and record it as the strategy identification result. The strategy identification result reflects the result with the highest probability of the combination of the category of the advertising delivery platform, the location of the advertising delivery, the time of the advertising delivery, and the mode of the advertising delivery. The strategy identification result includes the category of the advertising delivery platform, the location of the advertising delivery, the time of the advertising delivery, and the mode of the advertising delivery.

[0059] In one embodiment, in step S202, that is, performing material requirement identification on the requirement text to obtain a requirement identification result, includes:

[0060] Material factors are extracted from the demand text to obtain material factor results.

[0061] It can be understood that the process of extracting the material factors is as follows: first, the demand text is segmented and the demand text is split into the smallest unit characters or words to obtain multiple unit words; second, the split unit words are converted into word embedding vectors, and word vectors matching the unit words are searched from a preset corpus related to the material factors, and the word vectors found are used as the converted word vectors of the unit words; finally, the converted word vectors of each unit word are filtered, and the word vectors corresponding to the keywords related to the material factors are retained, and auxiliary words and entities irrelevant to the material factors are removed, and finally the entities or keywords corresponding to the remaining word vectors are used as the material factor results.

[0062] Based on the material factor result, the demand semantics of the demand text is identified to obtain a demand identification result corresponding to the material factor result.

[0063] It can be understood that the recognition of the demand semantics is a verification method that uses the Bi-LSTM algorithm, also known as the bidirectional long short-term memory network algorithm, to perform joint encoding in both forward and reverse directions to perform embedded word vector conversion to ensure the recognition process of converting into the most semantically consistent encoding, that is, the context of each unit word in the material factor result is predicted in both forward and reverse directions, so as to predict the most semantically consistent encoding corresponding to each unit word, and use the characters or words corresponding to each most semantically consistent encoding as the category result of a material factor in the demand recognition result.

[0064] The present invention achieves the goal of extracting material factors from the demand text to obtain material factor results; based on the material factor results, the demand semantics of the demand text are identified to obtain demand identification results corresponding to the material factor results. In this way, the demand semantics of each material factor can be automatically extracted through contextual demand semantics identification, thereby reducing the cost of manual identification and improving the accuracy and efficiency of material factor identification.

[0065] S203: Verify the requirement identification result according to the strategy identification result to obtain a material requirement result.

[0066] It can be understood that the verification is to weight the categories of the advertising delivery platform, the location of advertising delivery, the time of advertising delivery and the mode of advertising delivery in the strategy identification result to the categories of multiple advertising delivery platforms, the locations of multiple advertising delivery, the times of multiple advertising delivery and the modes of multiple advertising delivery in the demand identification result, so as to determine the category of an advertising delivery platform, the location of advertising delivery, the time of advertising delivery and the mode of advertising delivery according to each scoring value, and merge the determined category of an advertising delivery platform, the location of advertising delivery, the time of advertising delivery and the mode of advertising delivery with the category results related to the demand for advertising delivery or material, such as material size, material layout, material duration, etc. in the demand identification result, so as to obtain the material demand result, which reflects the category results of each material factor finally identified in the demand text.

[0067] The present invention achieves the following steps: performing text conversion on the demand file to obtain the demand text; performing material demand identification on the demand text to obtain a demand identification result; and performing material strategy identification on the demand text to obtain a strategy identification result; and verifying the demand identification result based on the strategy identification result to obtain a material demand result. In this way, the demand text of the text content can be automatically converted by using text conversion, and material demand identification and material strategy identification, as well as self-verification, can be used to obtain accurate and true material demand results, thereby ensuring the correctness of the material demand interpretation and improving the efficiency of the interpretation.

[0068] In one embodiment, step S20, i.e., performing material style recognition on the reference material to obtain a style recognition result, includes:

[0069] Perform image preprocessing on the reference material to obtain a preprocessed image.

[0070] It can be understood that the image preprocessing is the process of processing the reference material using an image enhancement algorithm. The image enhancement algorithm includes image denoising, increasing clarity (contrast), grayscale or obtaining image edge features or performing convolution, binarization, histogram equalization, Laplace transform and gamma transform on the image. The image enhancement algorithm is selected according to needs. For example: Laplace transform is performed on the reference material to enhance texture features, thereby obtaining the preprocessed image. The preprocessed image is an image of the reference material with enhanced texture features.

[0071] Style features are extracted from the preprocessed image to obtain a feature map.

[0072] It can be understood that the style feature extraction can be a parametric texture feature extraction method based on statistical distribution, or a non-parametric texture feature extraction method based on Markov random field, wherein the parametric texture feature extraction method based on statistical distribution is a statistical method that parameterizes the extracted texture by a statistical vector, that is, the texture feature extraction is modeled as an N-order statistic, and the model obtained by the modeling method of N-order texture feature learning, and the non-parametric texture feature extraction method based on Markov random field is to extract texture features using a Markov network, and the Markov network is a network that uses a small pixel point matrix of a preset size in the image to synthesize point by point with the similarity matching degree of various style matrices, and gradually learns similar texture features, and matrixes the extracted texture features to obtain the feature map, which reflects the texture features in the preprocessed image.

[0073] Multi-task material style recognition is performed on the feature map to identify the style recognition result.

[0074] It can be understood that the multi-task material style recognition is an identification process of performing probability prediction of each style category on the feature map through the branch network learned by each style, that is, the feature map is subjected to corresponding softmax processing through the fully connected layer learned by each style category to obtain the probability value corresponding to each branch network, and the style category corresponding to the probability value exceeding the preset threshold is recorded as the style recognition result. The style recognition result reflects the image style category of the reference material, for example: image style categories include aesthetic, retro, modern, deep and other style classifications.

[0075] The present invention achieves the following steps: performing image preprocessing on the reference material to obtain a preprocessed image; performing style feature extraction on the preprocessed image to obtain a feature map; performing multi-task material style recognition on the feature map to identify the style recognition result. In this way, the style of the reference material can be automatically identified, the workload of manual recognition is reduced, and accurate style labels are provided for the subsequent generation of material documents.

[0076] S30: Record the material requirement result and the style recognition result as a final requirement result.

[0077] Understandably, the requirement verification result and the style recognition result are formed into a one-dimensional array, and the one-dimensional array is recorded as the final requirement result, and the final requirement result reflects the requirements required by the material work order.

[0078] S40: Distribute the final demand result to the task queue corresponding to the material designer matching the final demand result based on the distribution rule.

[0079] It can be understood that the distribution rule is a rule for distribution based on the comprehensive score value of the historical scores of various material factors corresponding to each material designer. The distribution rule is constructed through a scoring model. The scoring model includes the average value of the historical scores of each material designer in various material factors, and clusters the historical scores of various material factors of each material designer. The clustering method is used to perform clustering processing, classify and identify the portrait label of each material designer, and the final demand result is matched with the average value of each material factor of each material designer through the scoring model, and the final demand result is matched with the portrait label corresponding to each material designer. The matching results of the two are combined to determine the matching degree of each material factor in the final demand result. For example, For example: obtain the average value corresponding to the material factors consistent with each item in the final demand result, and weight each obtained average value, for example, the weight of each material factor is one of the total number of all material factors, and increase the weight of the material factor corresponding to the portrait label, for example, increase the weight by 2 times on the basis of the original weight. A portrait label contains material factors that multiple material designers are good at. Finally, add the average values ​​corresponding to each material factor after adding weights, calculate the matching degree between each material designer and the final demand result, obtain the material designer corresponding to the maximum matching degree, record it as the material designer matching the final demand result, and distribute the final demand result to the task queue corresponding to the material designer.

[0080] Among them, the material designer is the staff who designs the material, the task queue is a collection of the task lists of the material designer, and the classification using the clustering method is to take the historical scores of the material factors corresponding to the material designer that are greater than the preset score value as a data set, and draw points from the data set. According to the degree of clustering of the drawn points, the material designer is determined to be a category of the corresponding portrait label. A category of a portrait label maps multiple material factors that are good at, indicating the material factors that the material designer of the category of the portrait label is good at mapping. The preset score value can be set according to demand, such as 7 points.

[0081] S50, receiving the material document returned by the task queue; the material document is a document created by the material designer according to the final requirement result.

[0082] It can be understood that the material document is a document created by the material designer based on the final demand result, that is, the material created by the material designer based on the final demand result. The material document is returned according to the sequence number of the task queue and is associated with the sequence number of the task queue and the material document.

[0083] S60: Convert the material document into a transmission format corresponding to the material work order, and push the converted material document to the demander corresponding to the material work order.

[0084] It can be understood that the conversion processing is a process of performing corresponding size conversion and encryption according to the transmission format associated with the demand party corresponding to different material work orders. The size conversion is a conversion process of adjusting the size of the image or / video of the material document to the size of the transmission format associated with the corresponding demand party. The size conversion is to use image size adjustment technology to proportionally enlarge or reduce each frame of the image to meet the size requirements of the transmission format. The encryption algorithm can be selected according to the requirements of the demand party, such as the encryption algorithm is a symmetric encryption algorithm or an asymmetric encryption algorithm, etc., and the converted material document is pushed to the terminal corresponding to the demand party corresponding to the material work order, and the material document related to the advertising delivery returned by the material work order is displayed.

[0085] The material work order data processing method, device, computer equipment and storage medium provided by the present invention obtain the demand file and reference material corresponding to the material work order; perform text recognition and demand extraction on the demand file to obtain a material demand result, and perform material style recognition on the reference material to obtain a style recognition result; record the material demand result and the style recognition result as the final demand result; based on the distribution rule, distribute the final demand result to the task queue corresponding to the material designer matching the final demand result; receive the material document returned by the task queue; the material document is a document created by the material designer according to the final demand result; convert the material document into a transmission format corresponding to the material work order, and push the converted material document to the demander corresponding to the material work order. Therefore, the method realizes the automatic identification of the final demand result by using text recognition, demand extraction and material style recognition, and automatically distributes it to the task queue corresponding to the matching material designer, and automatically converts the received material document to send it to the demander. Therefore, the manual workload is greatly reduced, the customer satisfaction of the pushed material documents is improved, and the high matching of the designed material documents is guaranteed.

[0086] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0087] In one embodiment, a material work order data processing device is provided, which corresponds one-to-one to the material work order data processing method in the above embodiment. Figure 4As shown, the material work order data processing device includes an acquisition module 11, an identification module 12, a recording module 13, a distribution module 14, a receiving module 15 and a conversion module 16. The functional modules are described in detail as follows:

[0088] Acquisition module 11, used to obtain the requirement file and reference material corresponding to the material work order;

[0089] The recognition module 12 is used to perform text recognition and demand extraction on the demand document to obtain a material demand result, and to perform material style recognition on the reference material to obtain a style recognition result;

[0090] A recording module 13, configured to record the material demand result and the style recognition result as a final demand result;

[0091] A distribution module 14 is configured to distribute the final demand result to a task queue corresponding to a material designer matching the final demand result based on a distribution rule;

[0092] The receiving module 15 is configured to receive the material document returned by the task queue; the material document is a document created by the material designer according to the final requirement result;

[0093] The conversion module 16 is configured to convert the material document into a transmission format corresponding to the material work order, and push the converted material document to the demander corresponding to the material work order.

[0094] In one embodiment, if Figure 5 As shown, the identification module 12 includes:

[0095] The preprocessing unit 21 is used to perform image preprocessing on the reference material to obtain a preprocessed image;

[0096] An extraction unit 22 is used to extract style features from the pre-processed image to obtain a feature map;

[0097] The recognition unit 23 is configured to perform multi-task material style recognition on the feature map and identify the style recognition result.

[0098] For the specific definition of the material work order data processing device, please refer to the definition of the material work order data processing method above, and will not be repeated here. The various modules in the above-mentioned material work order data processing device can be implemented in whole or in part by software, hardware, or a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0099] In one embodiment, a computer device is provided. The computer device can be a client or a server. The internal structure diagram thereof can be as follows: Figure 6 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a readable storage medium and an internal memory. The readable storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the readable storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a material work order data processing method is implemented.

[0100] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the material work order data processing method in the above embodiment is implemented.

[0101] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the material work order data processing method in the above embodiment is implemented.

[0102] Those skilled in the art will appreciate that all or part of the processes in the above-described embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the above-described embodiments. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAM bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM).

[0103] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0104] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A material work order data processing method, characterized in that: include: Obtain the requirement documents and reference materials corresponding to the material work order; Performing text recognition and demand extraction on the demand document to obtain a material demand result, and performing material style recognition on the reference material to obtain a style recognition result; Recording the material demand result and the style recognition result as the final demand result; Based on the distribution rules, the final demand result is distributed to the task queue corresponding to the material designer matching the final demand result; Receive the material document returned by the task queue; the material document is a document created by the material designer according to the final demand result; Convert the material document into a transmission format corresponding to the material work order, and push the converted material document to the demander corresponding to the material work order; The text recognition and demand extraction of the demand document to obtain the material demand results includes: Performing text conversion on the demand document to obtain a demand text; Performing material requirement identification on the requirement text to obtain a requirement identification result, and performing material strategy identification on the requirement text to obtain a strategy identification result; According to the strategy identification result, the requirement identification result is verified to obtain a material requirement result; the material requirement result reflects the category result of each material factor finally identified in the requirement text; The step of performing text conversion on the requirement document to obtain the requirement text includes: Performing file format recognition on the demand file to identify the file format of the demand file; Performing text conversion on the demand document according to the file format to obtain the demand text; The performing material requirement identification on the requirement text to obtain a requirement identification result includes: Extracting material factors from the demand text to obtain material factor results; Based on the material factor result, the demand semantics of the demand text are recognized to obtain a demand recognition result corresponding to the material factor result; The performing material style recognition on the reference material to obtain a style recognition result includes: Performing image preprocessing on the reference material to obtain a preprocessed image; Performing style feature extraction on the preprocessed image to obtain a feature map; Multi-task material style recognition is performed on the feature map to identify the style recognition result; the style recognition result reflects the image style category of the reference material.

2. The material work order data processing method according to claim 1, wherein: The step of converting the demand document into a text format corresponding to the file format to obtain the demand text includes: When it is detected that the file format is an audio format, using speech recognition technology to convert the required file into speech text to obtain the required text; When it is detected that the file format is a video format, the audio extraction technology and speech recognition technology are used to convert the required file into a video format to obtain the required text; When it is detected that the file format is an image format, the OCR technology is used to extract text from the demand file to obtain the demand text.

3. A material work order data processing device, characterized in that: include: The acquisition module is used to obtain the requirement files and reference materials corresponding to the material work order; an identification module for performing text recognition and demand extraction on the demand document to obtain a material demand result, and performing material style recognition on the reference material to obtain a style recognition result; A recording module, configured to record the material demand result and the style recognition result as a final demand result; A distribution module, configured to distribute the final demand result to a task queue corresponding to a material designer matching the final demand result based on a distribution rule; A receiving module, configured to receive the material document returned by the task queue; the material document is a document created by the material designer according to the final demand result; A conversion module, configured to convert the material document into a transmission format corresponding to the material work order, and push the converted material document to the demander corresponding to the material work order; The text recognition and demand extraction of the demand document to obtain the material demand results includes: Performing text conversion on the demand document to obtain a demand text; Performing material requirement identification on the requirement text to obtain a requirement identification result, and performing material strategy identification on the requirement text to obtain a strategy identification result; According to the strategy identification result, the requirement identification result is verified to obtain a material requirement result; the material requirement result reflects the category result of each material factor finally identified in the requirement text; The step of performing text conversion on the requirement document to obtain the requirement text includes: Performing file format recognition on the demand file to identify the file format of the demand file; Performing text conversion on the demand document according to the file format to obtain the demand text; The performing material requirement identification on the requirement text to obtain a requirement identification result includes: Extracting material factors from the demand text to obtain material factor results; Based on the material factor result, the demand semantics of the demand text are recognized to obtain a demand recognition result corresponding to the material factor result; The identification module includes: A preprocessing unit, configured to perform image preprocessing on the reference material to obtain a preprocessed image; an extraction unit, configured to extract style features from the preprocessed image to obtain a feature map; The recognition unit is used to perform multi-task material style recognition on the feature map to identify the style recognition result; the style recognition result reflects the image style category of the reference material.

4. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the material work order data processing method according to any one of claims 1 to 2 is implemented.

5. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the material work order data processing method according to any one of claims 1 to 2 is implemented.

Images