Service order quality evaluation method and device, electronic equipment and storage medium

Abstract

The application provides a business form quality evaluation method and device, electronic equipment and a computer readable storage medium. The business form quality evaluation method comprises: obtaining a target image of a to-be-evaluated business form; determining at least one of an architecture complexity of the to-be-evaluated business form and a text filling quality of the to-be-evaluated business form based on the target image, the architecture complexity being determined based on at least one of a table complexity of the to-be-evaluated business form and a picture-text complexity of the to-be-evaluated business form; determining target index information of the to-be-evaluated business form based on at least one of the architecture complexity and the text filling quality, the target index information being used to indicate a quality of the to-be-evaluated business form; and evaluating the quality of the to-be-evaluated business form based on the target index information to obtain the quality of the to-be-evaluated business form. In the application, the quality of the business form can be effectively and quickly evaluated, and the quality evaluation efficiency of the business form is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a business form quality evaluation method and device, electronic equipment and computer readable storage medium. BACKGROUND

[0002] The perfect quality of a business form such as a bank business has important guiding significance for the workload statistics of staff and the work quality of staff. However, the content of the business form is different, and the prior art currently does not have a better method for evaluating the quality of the business form.

[0003] Therefore, how to effectively evaluate the quality of the business form with different content to improve the quality evaluation efficiency of the business form has become a problem to be solved. SUMMARY

[0004] The present application provides a business form quality evaluation method and device, electronic equipment and computer readable storage medium, which can effectively and quickly evaluate the quality of the business form and improve the quality evaluation efficiency of the business form.

[0005] In a first aspect, the present application provides a business form quality evaluation method, which comprises:

[0006] obtaining a target image of a business form to be evaluated;

[0007] determining at least one of an architecture complexity of the business form to be evaluated and a text filling quality of the business form to be evaluated based on the target image, wherein the architecture complexity is determined based on at least one of a table complexity of the business form to be evaluated and a picture-text complexity of the business form to be evaluated;

[0008] determining target index information of the business form to be evaluated based on at least one of the architecture complexity and the text filling quality, wherein the target index information is used to indicate the quality of the business form to be evaluated;

[0009] evaluating the quality of the business form to be evaluated based on the target index information, to obtain the quality of the business form to be evaluated.

[0010] In a second aspect, the present application provides a business form quality evaluation device, which comprises:

[0011] an acquisition unit, configured to obtain a target image of a business form to be evaluated;

[0012] The identification unit is configured to determine at least one of an architecture complexity of the to-be-evaluated business form and a text filling quality of the to-be-evaluated business form based on the target image, wherein the architecture complexity is determined based on at least one of a table complexity of the to-be-evaluated business form and a graphic-text complexity of the to-be-evaluated business form;

[0013] The evaluation unit is configured to determine target index information of the to-be-evaluated business form based on at least one of the architecture complexity and the text filling quality, wherein the target index information is used to indicate a quality of the to-be-evaluated business form.

[0014] The evaluation unit is further configured to evaluate the quality of the to-be-evaluated business form based on the target index information, to obtain the quality of the to-be-evaluated business form.

[0015] In some embodiments, the target index information is the architecture complexity of the to-be-evaluated business form, and the identification unit is specifically configured to:

[0016] perform feature extraction on the target image through a feature extraction layer in a pre-trained architecture identification model, to obtain image features of the target image;

[0017] determine architecture information of the to-be-evaluated business form according to the image features through an information identification layer in the architecture identification model;

[0018] determine the architecture complexity of the to-be-evaluated business form based on the architecture information.

[0019] In some embodiments, the architecture information includes table information and graphic-text information of the to-be-evaluated business form, and the identification unit is specifically configured to:

[0020] determine a table complexity of the to-be-evaluated business form according to the table information of the to-be-evaluated business form;

[0021] determine a graphic-text complexity of the to-be-evaluated business form according to the graphic-text information of the to-be-evaluated business form;

[0022] determine the architecture complexity of the to-be-evaluated business form based on the table complexity and the graphic-text complexity.

[0023] In some embodiments, the target index information is the text filling quality of the to-be-evaluated business form, and the identification unit is specifically configured to:

[0024] perform text recognition based on the target image through a pre-trained text recognition model, to determine target text information of the to-be-evaluated business form;

[0025] The language fluency of the to-be-evaluated business form is identified based on the target text information through a pre-trained language model.

[0026] The number of filled words and the number of filled paragraphs of the to-be-evaluated business form are determined based on the target text information.

[0027] The text filling quality of the to-be-evaluated business form is determined based on the language fluency, the number of filled words, and the number of filled paragraphs.

[0028] In some embodiments, the identification unit is specifically configured to:

[0029] The target image is cut into a plurality of sub-region images of the target image through an image cutting layer in the text recognition model, where each sub-region image is an image of a region where a minimum unit of text is located.

[0030] Text information of each sub-region image is obtained through a text recognition layer in the text recognition model.

[0031] Target text information of the to-be-evaluated business form is obtained by fusing text information of a plurality of sub-region images through a text fusion layer in the text recognition model.

[0032] In some embodiments, the target indicator information includes an architecture complexity and a text filling quality of the to-be-evaluated business form, and the evaluation unit is specifically configured to:

[0033] The business type of the to-be-evaluated business form is obtained.

[0034] A first quality weight of the architecture complexity associated with the business type is obtained.

[0035] A second quality weight of the text filling quality associated with the business type is obtained.

[0036] The quality of the to-be-evaluated business form is determined according to the first quality weight, the second quality weight, the architecture complexity, and the text filling quality.

[0037] In some embodiments, the business form quality evaluation device further includes a scoring unit, which is specifically configured to:

[0038] The business type of the to-be-evaluated business form is obtained.

[0039] The business score of the operation object of the to-be-evaluated business form is determined based on the quality of the to-be-evaluated business form and the business type of the to-be-evaluated business form.

[0040] In a third aspect, the present application also provides an electronic device, comprising a processor and a memory, wherein the memory stores a computer program, and the processor invokes the computer program in the memory to execute the steps of any of the service order quality evaluation methods provided in the present application.

[0041] In a fourth aspect, the present application also provides a computer readable storage medium, which stores a computer program, and the computer program is loaded by a processor to execute the steps of the service order quality evaluation method.

[0042] In the present application, the target image of the service order to be evaluated is obtained, the target image is identified to obtain the target index information of the service order to be evaluated, such as at least one of the architecture complexity of the service order to be evaluated and the text filling quality of the service order to be evaluated, and the quality of the service order to be evaluated is evaluated based on the target index information to obtain the quality of the service order to be evaluated. Since the target index information that can be used to indicate the quality of the service order to be evaluated can be automatically extracted from the target image of the service order to be evaluated, the quality of the service order to be evaluated is evaluated, so that the quality of the service order can be effectively and quickly evaluated, and the quality evaluation efficiency of the service order is improved. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0044] Figure 1 is a scene schematic diagram of the service order quality evaluation system provided by the embodiments of the present application;

[0045] Figure 2 is a flowchart of the service order quality evaluation method provided by the embodiments of the present application;

[0046] Figure 3 is a network structure schematic diagram of the architecture identification model provided in the embodiments of the present application;

[0047] Figure 4 is another network structure schematic diagram of the architecture identification model provided in the embodiments of the present application;

[0048] Figure 5 is a network structure schematic diagram of the text identification model provided in the embodiments of the present application;

[0049] Figure 6 is an embodiment structure schematic diagram of the service order quality evaluation device provided in the embodiments of the present application;

[0050] Figure 7 FIG. 1 is a schematic structural diagram of an embodiment of an electronic device provided in the embodiments of the present application. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, any person skilled in the art can obtain all other embodiments without creative work, which are within the scope of protection of the present application.

[0052] In the description of the embodiments of the present application, it should be understood that the terms "first", "second" are used only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the embodiments of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0053] The following description is presented to enable any person skilled in the art to practice and use the present application. In the following description, details are set forth for the purpose of explanation. It will be appreciated that one of ordinary skill in the art can realize and implement the present application without using these specific details. In other instances, well-known processes have not been described in detail so as not to obscure the description of the embodiments of the present application. Accordingly, the present application is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

[0054] The execution subject of the service order quality evaluation method in the embodiments of the present application can be a service order quality evaluation device provided in the embodiments of the present application, or a server device, a physical host, or a user equipment (UE) and other types of electronic devices integrated with the service order quality evaluation device, wherein the service order quality evaluation device can be realized in the form of hardware or software, and the UE can be a terminal device such as a smart phone, a tablet computer, a notebook computer, a palm computer, a desktop computer, or a personal digital assistant (PDA).

[0055] The electronic device can adopt a separate running mode or a device cluster mode.

[0056] Referring to Figure 1 , Figure 1is a scenario diagram of a business form quality evaluation system provided by an embodiment of the present application. The business form quality evaluation system can include an electronic device 700, and the electronic device 700 can be integrated with a business form quality evaluation apparatus. For example, the electronic device can obtain a target image of a business form to be evaluated; determine at least one of an architecture complexity of the business form to be evaluated and a text filling quality of the business form to be evaluated based on the target image, wherein the architecture complexity is determined based on at least one of a table complexity of the business form to be evaluated and a picture-text complexity of the business form to be evaluated; determine target index information of the business form to be evaluated based on at least one of the architecture complexity and the text filling quality, wherein the target index information is used to indicate a quality of the business form to be evaluated; and evaluate the quality of the business form to be evaluated based on the target index information, to obtain the quality of the business form to be evaluated.

[0057] In addition, as shown in Figure 1 the business form quality evaluation system can further include a memory 200 configured to store data, such as objective index data.

[0058] It should be noted that Figure 1 the scenario diagram of the business form quality evaluation system shown in FIG. 1 is only an example, and the business form quality evaluation system and the scenario described in the embodiments of the present application are used to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, with the evolution of the business form quality evaluation system and the appearance of new business scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0059] Next, a business form quality evaluation method provided by an embodiment of the present application will be introduced. In the embodiments of the present application, an electronic device is used as an execution subject, and the execution subject will be omitted in subsequent method embodiments for simplification and convenience of description.

[0060] Referring to Figure 2 , Figure 2 is a flowchart of a business form quality evaluation method provided by an embodiment of the present application. It should be noted that, although a logical sequence is shown in the flowchart shown in Figure 2 or other drawings, in some cases, the steps shown or described can be performed in an order different from that shown here. The business form quality evaluation method includes steps 201-204, wherein:

[0061] 201, obtaining a target image of a business form to be evaluated.

[0062] The to-be-evaluated business form refers to a business form that needs to be evaluated, such as a bank business form. The to-be-evaluated business form can be an electronic business form or a paper business form, and the specific form of the to-be-evaluated business form is not limited herein.

[0063] The target image refers to an image of the to-be-evaluated business form.

[0064] In step 201, the target image of the to-be-evaluated business form can be obtained in various ways, for example, including:

[0065] (1) When the to-be-evaluated business form is an electronic business form, a screenshot of the to-be-evaluated business form is obtained as the target image.

[0066] (2) In actual application, the electronic device can be integrated with a camera on the hardware, and a video frame or an image of the business form placement area is obtained in real time by the camera, which is used as the target image.

[0067] (3) A camera can also be arranged above the business form placement area, and a video frame or an image of the business form placement area is obtained in real time by the camera above the business form placement area. The electronic device is connected to the camera above the business form placement area through a network, and the video frame or the image of the business form placement area obtained by the camera above the business form placement area is obtained online from the camera above the business form placement area, which is used as the target image.

[0068] (4) The electronic device can also read the image of the business form placement area obtained by the camera from a related storage medium that stores the image of the business form placement area obtained by the camera, which is used as the target image.

[0069] (5) The video frame or the image of the business form placement area pre-collected and stored in the electronic device is read as the target image.

[0070] The camera can obtain an image according to a preset shooting mode, for example, the shooting height, the shooting direction, or the shooting distance can be set, and the specific shooting mode can be adjusted according to the camera itself, which is not limited herein.

[0071] The method of obtaining the target image is only an example, and is not limited thereto.

[0072] 202, based on the target image, at least one of the architecture complexity of the to-be-evaluated business form and the text filling quality of the to-be-evaluated business form is determined.

[0073] The architecture complexity is determined based on at least one of a table complexity of the to-be-evaluated business form and a picture-text complexity of the to-be-evaluated business form.

[0074] 203. Determine target index information of the to-be-evaluated business form based on at least one of the architecture complexity and the text filling quality.

[0075] The target index information is index information used for evaluating the quality of the to-be-evaluated business form, and is used for indicating the quality of the to-be-evaluated business form, such as the architecture complexity and the text filling quality of the to-be-evaluated business form.

[0076] The architecture complexity is the complexity of the content of the to-be-evaluated business form determined based on architecture information such as table information and picture-text information of the to-be-evaluated business form, such as table complexity and picture-text complexity. The table information refers to table-related information in the to-be-evaluated business form, and can include, for example, the number of tables, the number of dimensions filled in each table, and the difficulty of filling content in each dimension. The picture-text information refers to the content of the combination of pictures (including pictures taken, data charts, etc.) and text in the to-be-evaluated business form, for example, pictures in the business form and text in the part of the description combined with the pictures.

[0077] The text filling quality is the filling quality of the content of the to-be-evaluated business form determined based on information such as language fluency, the number of filled words, and the number of filled paragraphs of the to-be-evaluated business form.

[0078] The following describes the identification process of the target index information of the to-be-evaluated business form, taking the target index information as the architecture complexity and the text filling quality, respectively.

[0079] I. The target index information is the architecture complexity.

[0080] Since the architecture complexity of the to-be-evaluated business form can reflect the complexity and workload of the improvement of the to-be-evaluated business form to a certain extent, the architecture complexity is used as an evaluation index of the quality of the to-be-evaluated business form, which reflects the quality of the to-be-evaluated business form to a certain extent, and further improves the objectivity of the evaluation of the quality of the to-be-evaluated business form.

[0081] For example, the architecture information in the to-be-evaluated business form can be identified by using a pre-trained architecture identification model, and the architecture complexity of the to-be-evaluated business form is determined based on the identified architecture information. At this time, step 202 can specifically include steps 2021A-2023A.

[0082] For the convenience of understanding, the network architecture and working principle of the architecture identification model are introduced as follows. Figure 3 As shown in FIG. 2, the architecture identification model includes a feature extraction layer and an information identification layer.

[0083] The feature extraction layer is configured to perform feature extraction on the image to obtain image features. The feature extraction layer can include a plurality of convolution layers and pooling layers, which are configured to perform convolution and pooling operations on the image. For example, a backbone network of a model that can be used for image classification, image detection, or image segmentation tasks can be extracted as the feature extraction layer of the architecture recognition model. For example, a backbone network of GoogleNet or ResNet can be extracted as the feature extraction layer of the architecture recognition model. The feature extraction layer takes the image as input, performs convolution, pooling, and other operations on the image, and outputs the image features.

[0084] The information recognition layer is configured to recognize the architecture information of the business form in the image based on the image features.

[0085] For example, the architecture recognition model can be trained by the following steps:

[0086] 1. Construct a preliminary architecture recognition model.

[0087] For example, an open-source network (such as a YOLOv network) with default model parameters (which can be used for detection tasks) can be used as the preliminary architecture recognition model. Similar to the trained architecture recognition model, the preliminary architecture recognition model can include a feature extraction layer and an information recognition layer. The feature extraction layer is configured to perform feature extraction on a sample image to obtain image features of the sample image, and the information recognition layer is configured to perform prediction based on the image features of the sample image to obtain architecture information of a business form in the sample image.

[0088] 2. Obtain a training data set.

[0089] The training data set includes a plurality of sample images, some of which include architecture information and some of which do not include architecture information. Each sample image in the training data set is labeled, and the label information includes an actual region detection frame and a category of the architecture information.

[0090] 3. Train the preliminary architecture recognition model using the training data set with the label information of the sample image as supervision until the preliminary architecture recognition model converges, thereby obtaining a trained architecture recognition model. At this time, the trained architecture recognition model can be applied to detect the architecture information in the image.

[0091] The trained architecture recognition model can learn the relationship between the architecture information and the image features, thereby accurately detecting the architecture information in the image.

[0092] 2021A. Perform feature extraction on the target image through the feature extraction layer of the pre-trained architecture recognition model to obtain image features of the target image.

[0093] 2022A, identifying the architecture information of the to-be-evaluated business order according to the image features through the information identification layer in the architecture identification model.

[0094] 2023A, determining the architecture complexity of the to-be-evaluated business order based on the architecture information.

[0095] For example, taking the architecture information as table information and the architecture complexity as table complexity as an example, first, in step 2021A, the target image is input into the pre-trained architecture identification, so as to perform feature extraction processing such as convolution operation and pooling operation on the target image through the feature extraction layer in the architecture identification model, to obtain the image features of the target image. Then, in step 2022A, the table information of the to-be-evaluated business order is obtained by predicting the image features of the target image through the information identification layer in the architecture identification model. Finally, in step 2023A, the table complexity of the to-be-evaluated business order is determined based on the table information of the to-be-evaluated business order according to the number of tables contained in the to-be-evaluated business order, the number of dimensions required to fill each table, the filling content difficulty of each dimension, etc., as the architecture complexity of the to-be-evaluated business order.

[0096] For example, taking the architecture information as table information and the architecture complexity as table complexity as an example, first, in step 2021A, the target image is input into the pre-trained architecture identification, so as to perform feature extraction processing such as convolution operation and pooling operation on the target image through the feature extraction layer in the architecture identification model, to obtain the image features of the target image. Then, in step 2022A, the table information of the to-be-evaluated business order is obtained by predicting the image features of the target image through the information identification layer in the architecture identification model. Finally, in step 2023A, the table complexity of the to-be-evaluated business order is determined based on the table information of the to-be-evaluated business order according to the number of tables contained in the to-be-evaluated business order, the number of dimensions required to fill each table, the filling content difficulty of each dimension, etc., as the architecture complexity of the to-be-evaluated business order.

[0097] For example, taking the architecture information as table information and the architecture complexity as table complexity as an example, first, in step 2021A, the target image is input into the pre-trained architecture identification, so as to perform feature extraction processing such as convolution operation and pooling operation on the target image through the feature extraction layer in the architecture identification model, to obtain the image features of the target image. Then, in step 2022A, the table information of the to-be-evaluated business order is obtained by predicting the image features of the target image through the information identification layer in the architecture identification model. Finally, in step 2023A, the table complexity of the to-be-evaluated business order is determined based on the table information of the to-be-evaluated business order according to the number of tables contained in the to-be-evaluated business order, the number of dimensions required to fill each table, the filling content difficulty of each dimension, etc., as the architecture complexity of the to-be-evaluated business order. Figure 4

[0098] In step 2021A, the first sub-extraction layer is used to perform feature extraction processing such as convolution operation and pooling operation on the target image to obtain the first image features of the target image. The first image features are used to predict the table information of the to-be-evaluated order.

[0099] ​In step 2021A, the second sub-extraction layer is configured to perform feature extraction processing such as convolution operation, pooling operation, etc. on the target image to obtain second image features of the target image. The second image features are used to predict the graphic-text information of the to-be-evaluated order.

[0100] In step 2022A, the first sub-recognition layer is configured to perform prediction based on the first image features of the target image to obtain the table information of the to-be-evaluated business order.

[0101] In step 2022A, the second sub-recognition layer is configured to perform prediction based on the second image features of the target image to obtain the graphic-text information of the to-be-evaluated business order.

[0102] In step 2023A, the table complexity of the to-be-evaluated business order is determined based on the table information of the to-be-evaluated business order; the graphic-text complexity of the to-be-evaluated business order is determined based on the graphic-text information of the to-be-evaluated business order; and the architecture complexity of the to-be-evaluated business order is determined based on the table complexity and the graphic-text complexity. The determination of the table complexity and the graphic-text complexity can refer to the related description above, which will not be repeated here. For example, the architecture complexity of the to-be-evaluated business order can be obtained by adding the table complexity and the graphic-text complexity with a certain weight ratio.

[0103] As can be seen, through steps 2021A-2023A, the architecture complexity of the to-be-evaluated business order can be extracted based on the table information and the graphic-text information which can reflect the complexity and quality of the to-be-evaluated business order, and used as an evaluation index of the quality of the to-be-evaluated business order, thereby improving the reliability of the quality of the to-be-evaluated business order. Moreover, since the pre-trained architecture recognition model has learned the relationship between the architecture information and the image features, the architecture complexity can be accurately detected by using the pre-trained architecture recognition model to recognize the architecture complexity.

[0104] II. The target index information is the quality of text filling.

[0105] Since the quality of text filling of the to-be-evaluated business order can reflect the complexity and workload of the to-be-evaluated business order to some extent, the quality of text filling can be used as an evaluation index of the quality of the to-be-evaluated business order, which can reflect the quality of the to-be-evaluated business order to some extent, thereby improving the objectivity of the evaluation of the quality of the to-be-evaluated business order.

[0106] For example, the language fluency, the number of filled words, and the number of filled paragraphs, etc. of the to-be-evaluated business order can be identified through the target image, which are used to determine the quality of text filling of the to-be-evaluated business order. At this time, step 202 can include steps 2021B-2024B as follows.

[0107] 2021B, performing text recognition on the target image based on the pre-trained text recognition model to determine target text information of the to-be-evaluated service order.

[0108] For the convenience of understanding, the network architecture and working principle of the text recognition model will be introduced first. As shown in Figure 5 The text recognition model includes an image segmentation layer, a text recognition layer, and a text fusion layer.

[0109] The image segmentation layer is used to detect the text of the target image and determine the area of each character or word. The target image is then segmented according to the area of each character or word to obtain multiple sub-region images of the target image. Each sub-region image only contains the smallest unit of text.

[0110] The text recognition layer is used to extract and recognize the text of each sub-region image based on the image features of each sub-region image to obtain the text information of each sub-region image.

[0111] The text fusion layer is used to extract and predict the text order of each sub-region image based on the image features of each sub-region image to obtain the text order of each sub-region image. The text information of each sub-region image is then sequentially spliced according to the text order of each sub-region image to obtain the target text information of the to-be-evaluated service order in the target image.

[0112] For example, the text recognition model can be trained by the following steps:

[0113] 1. Construct a preliminary text recognition model.

[0114] For example, an open-source network (such as YOLOv network) with default model parameters (which can be used for detection tasks) can be used as a preliminary text recognition model. Similar to the trained text recognition model, the preliminary text recognition model can include an image segmentation layer, a text recognition layer, and a text fusion layer. The image segmentation layer is used to detect the text of the sample image of the sample service order and determine the area of each character or word. The sample image is then segmented according to the area of each character or word to obtain multiple sub-region images of the sample image. The text recognition layer is used to recognize each sub-region image to obtain the predicted text information of each sub-region image. The text fusion layer is used to extract and predict the text order of each segmented image based on the image features of each segmented image to obtain the text prediction order of each segmented image. The predicted text information of each sub-region image is then sequentially spliced according to the text prediction order of each segmented image to obtain the sample text information of the sample service order.

[0115] 2. Obtain the training data set.

[0116] The training data set contains a plurality of sample images, and each sample image is an image containing text information of a sample service order. Each sample image in the training data set is labeled, and the label information can include a segmented image of each minimum unit text in the sample service order, actual text information in each segmented image, and actual text sorting of each segmented image.

[0117] 3. The training data set is used to train the preliminary text recognition model with the label information of the sample image as supervision until the preliminary text recognition model converges, and a trained text recognition model is obtained. At this time, the trained text recognition model can be applied to detect the text information in the image.

[0118] For example, the training process can include the following steps ①-②:

[0119] Step 1: First, input the sample image into the preliminary text recognition model. Then, perform character detection on the sample image of the sample service order through the image segmentation layer to determine the region of each character or each word, and segment the sample image according to the region of each character or each word to obtain a plurality of sub-region images of the sample image. Next, calculate the loss between each sub-region image and each segmented image as the segmentation loss of the image segmentation layer. Finally, adjust the model parameters of the image segmentation layer in the preliminary text recognition model according to the segmentation loss of the image segmentation layer, and when the first preset condition is met, the text recognition model with the adjusted model parameters of the image segmentation layer is used as the first recognition model. Step 2.

[0120] Step 2: First, extract the text information of each segmented image through the text recognition layer in the first recognition model and perform text recognition on each segmented image according to the image features of each segmented image to obtain the predicted text information of each segmented image. Then, calculate the loss between the predicted text information of each segmented image and the actual text information as the recognition loss of the text recognition layer. Finally, adjust the model parameters of the text recognition layer in the first recognition model according to the recognition loss of the text recognition layer, and when the second preset condition is met, the text recognition model with the adjusted model parameters of the text recognition layer is used as the second recognition model. Step 3.

[0121] Step 3: First, the text fusion layer in the second recognition model is used to extract and predict the text order of each segmented image according to the image features of each segmented image, and the text prediction order of each segmented image is obtained. Then, the loss between the text prediction order of each segmented image and the actual text order is calculated as the fusion loss of the text fusion layer. Finally, the model parameters of the text fusion layer in the second recognition model are adjusted according to the fusion loss of the text fusion layer, and when the third preset condition is met, the model parameters of the text fusion layer after adjustment are used as the trained text recognition model.

[0122] As can be seen, in the first aspect, since the image segmentation layer in the trained text recognition model can fully learn the relationship between the smallest unit of text and the image features of each region, the image segmentation layer can accurately segment the image containing only the region where the smallest unit of text is located, thereby improving the recognition accuracy of the text information. In the second aspect, since the text recognition layer in the trained text recognition model can fully learn the relationship between the text information and the image features of the sub-region image, the text recognition layer can accurately recognize the text information in each sub-region image, thereby improving the recognition accuracy of the text information. In the third aspect, since the text fusion layer in the trained text recognition model can fully learn the relationship between the text order of each sub-region and the image features of each sub-region image, the text fusion layer can accurately restore the text order in each sub-region image, improve the fusion accuracy of the text information of each sub-region image, and thereby improve the recognition accuracy of the text information.

[0123] At this time, in step 2021B, the image segmentation layer in the trained text recognition model can be used to segment the target image to obtain a plurality of sub-region images of the target image; the text recognition layer in the trained text recognition model can be used to recognize the text information of each sub-region image; and the text fusion layer in the trained text recognition model can be used to fuse the text information of the plurality of sub-region images to obtain the target text information of the to-be-evaluated business form. Each sub-region image is an image of a region where a smallest unit of text is located.

[0124] 2022B, based on the target text information, the language fluency of the to-be-evaluated business form is identified by using a pre-trained language model.

[0125] For example, first, the target text information is segmented and represented as a vector form to obtain a representation vector of the target text information. The representation vector of the target text information is input into the pre-trained language model for fluency prediction, and the language fluency of the to-be-evaluated business form is output.

[0126] 2023B, determine the filling word number and the filling paragraph number of the to-be-evaluated service order based on the target text information.

[0127] Specifically, the filling word number and the filling paragraph number of the to-be-evaluated service order can be obtained by directly counting the target text information.

[0128] 2024B, determine the text filling quality of the to-be-evaluated service order based on the language fluency, the filling word number and the filling paragraph number.

[0129] Exemplarily, the language fluency, the filling word number and the filling paragraph number all have a positive relationship with the text filling quality, that is, the higher the language fluency, the higher the text filling quality of the to-be-evaluated service order; the more the filling word number, the higher the text filling quality of the to-be-evaluated service order; the more the filling paragraph number, the higher the text filling quality of the to-be-evaluated service order.

[0130] It can be seen that, in the first aspect, the text filling quality of the to-be-evaluated service order can be extracted as an evaluation index of the quality of the to-be-evaluated service order based on the language fluency, the filling word number and the filling paragraph number, which can reflect the complexity and quality of the to-be-evaluated service order, so as to improve the reliability of the quality of the to-be-evaluated service order. In the second aspect, since the pre-trained text recognition model sufficiently learns the relationship between text information and image features, the target text information can be accurately detected by the pre-trained text recognition model.

[0131] 204, evaluate the quality of the to-be-evaluated service order based on the target index information, and obtain the quality of the to-be-evaluated service order.

[0132] In some embodiments, the target index information is the architecture complexity, which can be directly multiplied by a preset first weight coefficient as the quality of the to-be-evaluated service order.

[0133] In some embodiments, the target index information is the text filling quality, which can be directly multiplied by a preset second weight coefficient as the quality of the to-be-evaluated service order.

[0134] In some embodiments, the target indicator information includes the architecture complexity and the text filling quality, and a first result of multiplying the architecture complexity by a preset first weight coefficient and a second result of multiplying the text filling quality by a preset second weight coefficient are added according to a certain weight ratio as the quality of the business form to be evaluated. Further, different weights can be set for the architecture complexity and the text filling quality in combination with different business types of the business form, so as to obtain the quality of the business form to be evaluated according to different weight ratios, to meet the needs of different business scenarios. That is, step 204 can specifically include: obtaining a business type of the business form to be evaluated; obtaining a first quality weight of the architecture complexity associated with the business type; obtaining a second quality weight of the text filling quality associated with the business type; and determining the quality of the business form to be evaluated according to the first quality weight, the second quality weight, the architecture complexity and the text filling quality.

[0135] Further, in order to give a reasonable business score to an operation object of the business form (such as a perfect business employee of the business form), the quality of the business form to be evaluated can also be used to determine the business score of the operation object of the business form to be evaluated. Further, different business scores can be given to the same quality of the business form in combination with different business types, that is, the business form quality evaluation method further includes: obtaining a business type of the business form to be evaluated; and determining a business score of an operation object of the business form to be evaluated based on the quality of the business form to be evaluated and the business type of the business form to be evaluated. To meet the needs of business scenarios and realize automatic change of specific measurement standards of the business score according to changes of the business scenarios.

[0136] As can be seen from the above, by obtaining a target image of a business form to be evaluated; identifying the target image to obtain target indicator information of the business form to be evaluated; and evaluating the quality of the business form to be evaluated based on the target indicator information, the quality of the business form to be evaluated is obtained. Since the target indicator information that can be used to indicate the quality of the business form to be evaluated can be automatically extracted from the target image of the business form to be evaluated, the quality of the business form to be evaluated is evaluated, so that the quality of the business form can be effectively and quickly evaluated, and the quality evaluation efficiency of the business form is improved.

[0137] In order to better implement the above method, an embodiment of the present application further provides a business form quality evaluation device, which can be integrated in an electronic device, such as a computer device, which can be a terminal, a server or the like.

[0138] The terminal can be a mobile phone, a tablet computer, a smart Bluetooth device, a notebook computer, a personal computer or the like, and the server can be a single server or a server cluster composed of multiple servers.

[0139] For example, in the embodiment, the business form quality evaluation device is specifically integrated in a computer, and the method of the embodiment is described in detail.

[0140] For example, as shown in Figure 6 The business form quality evaluation device 600 can include:

[0141] An acquisition unit 601, configured to acquire a target image of a business form to be evaluated;

[0142] An identification unit 602, configured to determine at least one of an architecture complexity of the business form to be evaluated and a text filling quality of the business form to be evaluated based on the target image, wherein the architecture complexity is determined based on at least one of a table complexity of the business form to be evaluated and a picture-text complexity of the business form to be evaluated;

[0143] The evaluation unit 603 is further configured to determine target index information of the business form to be evaluated based on at least one of the architecture complexity and the text filling quality;

[0144] An evaluation unit 603, configured to evaluate a quality of the business form to be evaluated based on the target index information, to obtain the quality of the business form to be evaluated.

[0145] In some embodiments, the target index information is the architecture complexity of the business form to be evaluated, and the identification unit 602 is specifically configured to:

[0146] extract features of the target image through a feature extraction layer in a pre-trained architecture identification model, to obtain image features of the target image;

[0147] determine architecture information of the business form to be evaluated according to the image features through an information identification layer in the architecture identification model;

[0148] determine the architecture complexity of the business form to be evaluated based on the architecture information.

[0149] In some embodiments, the architecture information includes table information and picture-text information of the business form to be evaluated, and the identification unit 602 is specifically configured to:

[0150] determine a table complexity of the business form to be evaluated according to the table information of the business form to be evaluated;

[0151] determine a picture-text complexity of the business form to be evaluated according to the picture-text information of the business form to be evaluated;

[0152] determine the architecture complexity of the business form to be evaluated based on the table complexity and the picture-text complexity.

[0153] In some embodiments, the target index information is a text filling quality of the to-be-evaluated business form, and the identification unit 602 is specifically configured to:

[0154] perform text recognition on the target image based on the pre-trained text recognition model to determine target text information of the to-be-evaluated business form;

[0155] identify language fluency of the to-be-evaluated business form based on the target text information by using a pre-trained language model;

[0156] determine a number of filled words and a number of filled paragraphs of the to-be-evaluated business form based on the target text information;

[0157] determine the text filling quality of the to-be-evaluated business form based on the language fluency, the number of filled words, and the number of filled paragraphs.

[0158] In some embodiments, the identification unit 602 is specifically configured to:

[0159] perform image segmentation on the target image by using an image segmentation layer in the text recognition model to obtain a plurality of sub-region images of the target image, wherein each sub-region image is an image of a region in which a minimum unit of text is located;

[0160] perform text recognition on each sub-region image by using a text recognition layer in the text recognition model to obtain text information of each sub-region image;

[0161] fuse the text information of the plurality of sub-region images by using a text fusion layer in the text recognition model to obtain the target text information of the to-be-evaluated business form.

[0162] In some embodiments, the target index information includes an architecture complexity and a text filling quality of the to-be-evaluated business form, and the evaluation unit 603 is specifically configured to:

[0163] obtain a business type of the to-be-evaluated business form;

[0164] obtain a first quality weight of the architecture complexity associated with the business type;

[0165] obtain a second quality weight of the text filling quality associated with the business type;

[0166] determine a quality of the to-be-evaluated business form according to the first quality weight, the second quality weight, the architecture complexity, and the text filling quality.

[0167] In some embodiments, the business form quality evaluation apparatus 600 further includes a scoring unit (not shown in the figure), and the scoring unit is specifically configured to:

[0168] obtaining a service type of the to-be-evaluated service order;

[0169] determining a service score of an operation object of the to-be-evaluated service order based on the quality of the to-be-evaluated service order and the service type of the to-be-evaluated service order.

[0170] As can be seen, the service order quality evaluation device 600 of the embodiment can obtain a target image of a to-be-evaluated service order by the obtaining unit 601; determine at least one of an architecture complexity of the to-be-evaluated service order and a text filling quality of the to-be-evaluated service order based on the target image by the identifying unit 602, wherein the architecture complexity is determined based on at least one of a table complexity of the to-be-evaluated service order and a picture-text complexity of the to-be-evaluated service order; determine target index information of the to-be-evaluated service order based on at least one of the architecture complexity and the text filling quality by the evaluation unit 603; and evaluate the quality of the to-be-evaluated service order based on the target index information by the evaluation unit 603 to obtain the quality of the to-be-evaluated service order. Thus, the service order quality evaluation device 600 of the embodiment can automatically extract target index information that can be used to indicate the quality of the to-be-evaluated service order from the target image of the to-be-evaluated service order, evaluate the quality of the to-be-evaluated service order, and thus can effectively and quickly evaluate the quality of the service order and improve the quality evaluation efficiency of the service order.

[0171] Correspondingly, the embodiment of the present application further provides an electronic device. The electronic device can be a terminal, which can be a smart phone, a tablet computer, a notebook computer, a touch screen, a game console, a personal computer (PC), a personal digital assistant (PDA), and the like. As shown in Figure 7 Figure 7 The electronic device 700 includes a processor 701 having one or more processing cores, a memory 702 having one or more computer readable storage media, and a computer program stored on the memory 702 and executable on the processor. The processor 701 is electrically connected to the memory 702. Those skilled in the art can understand that the structure of the electronic device shown in the figure does not constitute a limitation on the electronic device, and can include more or fewer components than shown, or combine certain components, or different component arrangements.

[0172] ​The processor 701 is the control center of the electronic device 700, connects each part of the entire electronic device 700 through various interfaces and lines, executes various functions of the electronic device 700 and processes data by running or loading the software programs and / or modules stored in the memory 702 and calling the data stored in the memory 702, thereby overall monitoring the electronic device 700.

[0173] In the embodiment of the present application, the processor 701 in the electronic device 700 will load the instructions corresponding to the processes of one or more application programs into the memory 702, and run the application programs stored in the memory 702 by the processor 701, so as to realize various functions according to the following steps:

[0174] Obtain a target image of a to-be-evaluated business form;

[0175] Determine at least one of an architecture complexity of the to-be-evaluated business form and a text filling quality of the to-be-evaluated business form based on the target image, wherein the architecture complexity is determined based on at least one of a table complexity of the to-be-evaluated business form and a picture-text complexity of the to-be-evaluated business form;

[0176] Determine target index information of the to-be-evaluated business form based on at least one of the architecture complexity and the text filling quality, wherein the target index information is used to indicate a quality of the to-be-evaluated business form;

[0177] Evaluate the quality of the to-be-evaluated business form based on the target index information, and obtain the quality of the to-be-evaluated business form.

[0178] The specific implementation of each operation can refer to the previous embodiments, which will not be repeated here.

[0179] Optionally, as shown in Figure 7 The electronic device 700 further includes a touch display screen 703, a radio frequency circuit 704, an audio circuit 705, an input unit 706, and a power supply 707. The processor 701 is electrically connected with the touch display screen 703, the radio frequency circuit 704, the audio circuit 705, the input unit 706, and the power supply 707, respectively. Those skilled in the art can understand that the electronic device structure shown in Figure 7 The electronic device structure shown in the figure does not constitute a limitation on the electronic device, and can include more or fewer components than the figure, or combine certain components, or different component arrangements.

[0180] The touch display screen 703 can be used to display a graphical user interface and receive operation instructions generated by user acting on the graphical user interface. The touch display screen 703 can include a display panel and a touch panel. The display panel can be used to display information input by the user or provided to the user and various graphical user interfaces of the electronic device, which can be composed of graphics, text, icons, video and any combination thereof. Optionally, the display panel can be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), etc. The touch panel can be used to collect touch operations (such as user operations on or near the touch panel using a finger, a stylus or any suitable object or accessory) of the user thereon or therearound, and generate corresponding operation instructions, and the operation instructions execute corresponding programs. Optionally, the touch panel can include two parts of a touch detection device and a touch controller. The touch detection device detects the touch position of the user and detects the signals generated by the touch operation, and transmits the signals to the touch controller; the touch controller receives the touch information from the touch detection device, and converts it into touch coordinates, and then sends it to the processor 701, and can also receive commands from the processor 701 and execute them. The touch panel can cover the display panel, and when the touch panel detects a touch operation thereon or therearound, it transmits to the processor 701 to determine the type of the touch event, and then the processor 701 provides corresponding visual output on the display panel according to the type of the touch event. In the embodiments of the present application, the touch panel and the display panel can be integrated into the touch display screen 703 to realize the input and output functions. However, in some embodiments, the touch panel and the touch panel can realize the input and output functions as two independent components. That is, the touch display screen 703 can also realize the input function as part of the input unit 706.

[0181] The radio frequency circuit 704 can be used to transceive radio frequency signals to establish wireless communication with network devices or other electronic devices.

[0182] The audio circuit 705 can be used to provide an audio interface between the user and the electronic device through the speaker and the microphone. The audio circuit 705 can convert the received audio data into an electrical signal and transmit it to the speaker, which converts it into a sound signal output. On the other hand, the microphone collects sound signals and converts them into electrical signals, which are received by the audio circuit 705 and converted into audio data. After being processed by the processor 701, the audio data is transmitted to another electronic device through the radio frequency circuit 704, or output to the memory 702 for further processing. The audio circuit 705 can also include an earphone jack to provide communication between the external earphone and the electronic device.

[0183] The input unit 706 can be configured to receive inputted digital, character information or user feature information (e.g., fingerprint, iris, face information, etc.), and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0184] The power supply 707 is configured to supply power to various components of the electronic device 700. Optionally, the power supply 707 can be logically connected to the processor 701 through a power management system, so as to realize functions such as management of charging, discharging and power consumption management through the power management system. The power supply 707 can also include one or more direct current or alternating current power supplies, recharging systems, power failure detection circuits, power converters or inverters, power status indicators, etc.

[0185] Although Figure 7 The electronic device 700 can also include a camera, a sensor, a wireless fidelity module, a Bluetooth module, etc., which are not shown in the embodiment, and will not be described herein.

[0186] In the above embodiments, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0187] As can be seen from the above, the electronic device provided in the embodiment can bring the following technical effects: since the target indicator information that can be used to indicate the quality of the to-be-evaluated business form can be automatically extracted from the target image of the to-be-evaluated business form, and the quality of the to-be-evaluated business form is evaluated, the quality of the business form can be effectively and quickly evaluated, and the quality evaluation efficiency of the business form is improved.

[0188] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions, or by instructions controlling related hardware, which can be stored in a computer readable storage medium and loaded and executed by a processor.

[0189] To this end, the embodiment of the present application provides a computer readable storage medium, which stores a plurality of computer programs. The computer programs can be loaded by a processor to execute the steps in any of the business form quality evaluation methods provided by the embodiments of the present application. For example, the computer programs can execute the following steps:

[0190] Obtaining a target image of a to-be-evaluated business form;

[0191] Determining at least one of an architecture complexity of the to-be-evaluated business form and a character filling quality of the to-be-evaluated business form based on the target image, wherein the architecture complexity is determined based on at least one of a table complexity of the to-be-evaluated business form and a picture-text complexity of the to-be-evaluated business form.

[0192] determine target index information of the to-be-evaluated business form based on at least one of the architecture complexity and the text padding quality, wherein the target index information is used to indicate the quality of the to-be-evaluated business form;

[0193] evaluate the quality of the to-be-evaluated business form based on the target index information, to obtain the quality of the to-be-evaluated business form.

[0194] It can be seen that the computer program can be loaded by the processor to execute the steps in any of the business form quality evaluation methods provided by the embodiments of the present application, thereby bringing the following technical effects: since the target index information that can be used to indicate the quality of the to-be-evaluated business form can be automatically extracted from the target image of the to-be-evaluated business form, and the quality of the to-be-evaluated business form is evaluated, the quality of the business form can be effectively and quickly evaluated, and the quality evaluation efficiency of the business form is improved.

[0195] The specific implementation of each operation can be referred to the foregoing embodiments, which will not be described here.

[0196] The computer readable storage medium can include a read only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0197] Since the computer program stored in the computer readable storage medium can execute the steps in any of the business form quality evaluation methods provided by the embodiments of the present application, the beneficial effects of any of the business form quality evaluation methods provided by the embodiments of the present application can be achieved, which will be described in detail in the foregoing embodiments, and will not be described here.

[0198] The above describes in detail a business form quality evaluation method, device, electronic equipment and computer readable storage medium provided by the embodiments of the present application, and the principle and implementation manner of the present application are described by applying specific examples; the above embodiment description is only used to help understand the method and its core idea of the present application; meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manner and application range will be changed, and the above description should not be understood as limiting the present application.

Claims

1. A method of service quality assessment, characterized by, The method comprises: acquiring a target image of a to-be-evaluated business form; determining at least one of an architecture complexity of the to-be-evaluated business form and a text filling quality of the to-be-evaluated business form based on the target image, wherein the architecture complexity is determined based on at least one of a table complexity of the to-be-evaluated business form and a picture-text complexity of the to-be-evaluated business form; determining target index information of the to-be-evaluated business form based on at least one of the architecture complexity and the text filling quality, wherein the target index information is used to indicate a quality of the to-be-evaluated business form; evaluating the quality of the to-be-evaluated business form based on the target index information to obtain the quality of the to-be-evaluated business form.

2. The service quality evaluation method according to claim 1, characterized by, The target index information is the architecture complexity of the to-be-evaluated business form, and determining at least one of the architecture complexity of the to-be-evaluated business form, table information of the to-be-evaluated business form and picture-text information of the to-be-evaluated business form based on the target image comprises: performing feature extraction on the target image through a feature extraction layer in a pre-trained architecture recognition model to obtain image features of the target image; determining architecture information of the to-be-evaluated business form according to the image features through an information recognition layer in the architecture recognition model; determining the architecture complexity of the to-be-evaluated business form based on the architecture information.

3. The service quality evaluation method according to claim 2, characterized by, The architecture information comprises the table information and the picture-text information of the to-be-evaluated business form, and determining the architecture complexity of the to-be-evaluated business form based on the architecture information comprises: determining a table complexity of the to-be-evaluated business form according to the table information of the to-be-evaluated business form; determining a picture-text complexity of the to-be-evaluated business form according to the picture-text information of the to-be-evaluated business form; determining the architecture complexity of the to-be-evaluated business form based on the table complexity and the picture-text complexity.

4. The service quality evaluation method according to claim 1, characterized by, The target index information is the text filling quality of the to-be-evaluated business form, and determining at least one of the architecture complexity of the to-be-evaluated business form, the table information of the to-be-evaluated business form and the picture-text information of the to-be-evaluated business form based on the target image comprises: performing text recognition based on the target image through a pre-trained text recognition model to determine target text information of the to-be-evaluated business form; recognizing language fluency of the to-be-evaluated business form based on the target text information through a pre-trained language model; determining a filling word number and a filling paragraph number of the to-be-evaluated business form based on the target text information; determining the text filling quality of the to-be-evaluated business form based on the language fluency, the filling word number and the filling paragraph number.

5. The service quality evaluation method according to claim 4, characterized by, The performing text recognition based on the target image through the pre-trained text recognition model to determine the target text information of the to-be-evaluated business form comprises: performing segmentation on the target image through an image segmentation layer in the text recognition model to obtain a plurality of sub-region images of the target image, wherein each sub-region image is an image of a region where a minimum unit of text is located; text information of each of the sub-region images is obtained through a text recognition layer in the text recognition model; target text information of the to-be-evaluated business form is obtained by fusing the text information of the plurality of sub-region images through a text fusion layer in the text recognition model.

6. The service quality evaluation method according to claim 1, characterized by, The target index information includes an architecture complexity and a text filling quality of the to-be-evaluated business form. The quality of the to-be-evaluated business form is evaluated based on the target index information, and the quality of the to-be-evaluated business form is obtained, including: obtaining a business type of the to-be-evaluated business form; obtaining a first quality weight of the architecture complexity associated with the business type; obtaining a second quality weight of the text filling quality associated with the business type; determining the quality of the to-be-evaluated business form according to the first quality weight, the second quality weight, the architecture complexity, and the text filling quality.

7. The service quality evaluation method according to any one of claims 1 to 6, characterized by, The method further includes: obtaining a business type of the to-be-evaluated business form; determining a business score of an operation object of the to-be-evaluated business form based on the quality of the to-be-evaluated business form and the business type of the to-be-evaluated business form.

8. A service quality evaluation device characterized by comprising: The business form quality evaluation device includes: an obtaining unit configured to obtain a target image of a to-be-evaluated business form; an identifying unit configured to determine at least one of an architecture complexity of the to-be-evaluated business form and a text filling quality of the to-be-evaluated business form based on the target image, wherein the architecture complexity is determined based on at least one of a table complexity of the to-be-evaluated business form and a picture-text complexity of the to-be-evaluated business form; an evaluating unit configured to determine target index information of the to-be-evaluated business form based on at least one of the architecture complexity and the text filling quality, wherein the target index information is used to indicate a quality of the to-be-evaluated business form; The evaluating unit is further configured to evaluate the quality of the to-be-evaluated business form based on the target index information, and obtain the quality of the to-be-evaluated business form.

9. An electronic device, comprising: A processor and a memory are included, and the memory stores a computer program. When the processor invokes the computer program in the memory, the business form quality evaluation method according to any one of claims 1 to 7 is executed.

10. A computer-readable storage medium, characterized in that, A computer program is stored thereon, and the computer program is loaded by a processor to execute the steps in the business form quality evaluation method according to any one of claims 1 to 7.

Images