CN110879988B

CN110879988B - Method, client, device, server and computer readable medium for detecting information compliance

Info

Publication number: CN110879988B
Application number: CN201911153781.7A
Authority: CN
Inventors: 华绘
Original assignee: Weifang Jie'er Network Technology Co ltd
Current assignee: Weifang Jie'er Network Technology Co ltd
Priority date: 2019-11-22
Filing date: 2019-11-22
Publication date: 2024-06-11
Anticipated expiration: 2039-11-22
Also published as: CN110879988A

Abstract

The invention provides a method, a client, a device, a server and a computer readable medium for detecting information compliance, wherein the method comprises the following steps: receiving a picture and encoding the picture into picture information; detecting character information in the picture information by using a character detection model, determining whether the character information is compliant, and if the character information is not compliant, sending non-compliant information to a client so as to remind a user that the character is not compliant; and if so, obtaining the front character region picture, encoding the front character region picture into picture information, identifying the character picture information by using the identification information identification model, and determining whether uniform identification information in the character picture information is in compliance. The invention can detect and identify the picture by using the information compliance detection method, and then send the information which is judged to be non-compliance to the client. The method can monitor and judge whether staff wears uniforms according to regulations, maintains the brand image of enterprises, improves customer satisfaction, is more convenient for site management, has stronger model expansibility and higher application value.

Description

Method, client, device, server and computer readable medium for detecting information compliance

Technical Field

The present invention relates to the field of internet technologies, and in particular, to a method, a client, a device, a server, and a computer readable medium for detecting information compliance.

Background

With the rise of the concept of 'internet +', deep fusion and landing application are performed in a plurality of industries in recent years, such as O2O on-the-door service, intelligent school, intelligent park and the like, and the typical demands are to rationalize and control the staff thereof so as to achieve the effects of safe production and improving customer satisfaction. The uniform mark of the approach staff or the gate-up service staff not only represents the enterprise image, but also is convenient for field management and control, and the relevant responsibility units are positioned in time when accidents occur.

Before entering a relevant place, the entering staff generally performs fingerprint, face or card punching recognition mode at the entrance gate to verify the identity of the entering staff, and can not recognize and prompt uniform marks of the entering staff. The service personnel on the gate normally can self-shoot pictures after finishing to prove that the work is finished, the rear-end management and control personnel can conduct picture auditing, but the traditional auditing method still discovers whether uniform violations exist according to naked eyes, general clients cannot complain and do not conduct processing and responsibility following, the service personnel on the gate can be caused to have poor responsibility for a long time, even if the service personnel on the gate do not wear enterprise uniform or wear competitive product uniform, the service personnel on the gate cannot be punished, the enterprise image is lost, and the customer complaint rate is increased.

Disclosure of Invention

In view of the above, the present invention provides a method, a client, an apparatus, a server, and a computer-readable medium for information compliance detection, which further at least partially solve one or more of the problems due to the limitations and disadvantages of the related art.

A first aspect of the present invention provides a method for detecting information compliance, applied to a server, comprising the steps of:

receiving a picture from a worker client and encoding the picture into picture information, wherein the picture is an uploaded self-shot picture of the worker;

Detecting character information in the picture information by using a character detection model, determining whether the character information is qualified or not, wherein the character information comprises human body information containing the whole human face characteristics, and if the character information is not qualified, sending the non-qualified information to a client so as to remind a user that the character is not qualified, and restarting the first step;

And if the identification information is not in compliance, sending non-compliance information to a client so as to remind a user of the non-compliance of the uniform.

Further, after sending the non-compliance information to the client, the method further includes: responding to the user operation, if the worker selects to re-detect, receiving a new picture, and re-executing the first step; or the staff member chooses to end.

Further, the human detection model comprises a human detection model HF-ZF which is subjected to a large number of picture target labels and is trained.

Further, the identification information recognition model includes an identification classification model LRM that has undergone a large number of picture classifications and completed training.

Further, receiving the picture and encoding the picture into picture information, including performing base64 encoding processing on the picture to form the picture information.

Further, the person detection model is used for detecting the person information in the picture information to determine whether the picture information is compliant, and the method further comprises: inputting the picture information into a character detection model HF-ZF which has completed training; and carrying out compliance judgment on the model output result.

Further, inputting the picture information into the character detection model HF-ZF which has completed training, further includes: acquiring self-timer photos uploaded by all staff in a period of time; screening the pictures according to the principle that the human body with the complete face in the picture can be observed, so as to obtain a plurality of pictures meeting the condition; marking a frame of a human body containing a whole face in the picture and storing the frame as a data file in a corresponding format; based on a Faster-R-CNN model structure, in order to increase the influence of face features, in a picture feature convolution stage, firstly carrying out resolution to a fixed size M x N on a marked picture, respectively taking an upward quarter region and all picture regions after the fixed size as inputs, entering a first ZF convolution neural network and a second ZF convolution neural network, splicing a second convolution output feature map of the first ZF neural network and a fourth convolution output feature map of the second ZF neural network to finish feature fusion, taking the second convolution output feature map as input to carry out feature capture on a fifth convolution layer in the second convolution neural network, carrying out splicing and finishing feature fusion on a third convolution output feature map of the first ZF neural network and a fifth convolution output feature map of the second ZF neural network, and taking the third convolution output feature map as input to carry out feature capture on a sixth convolution layer in the second convolution neural network; substituting the marked data file into the adjusted model to train to obtain the character detection model HF-ZF.

Further, the method for judging compliance of the model output result comprises the following steps: setting a threshold value for the model result softmax probability output, and outputting only frame coordinate points larger than the threshold value; if the length of the output frame coordinate point set is 1, extracting a human body image containing human face features from the original picture according to the frame coordinate points; if the length of the output frame coordinate point set is not equal to 1, judging that the character is not in compliance, wherein the type of the character is not in compliance.

Further, the identification information identification model is used for identifying the picture information, and whether uniform identification information in the picture information is compliant is determined, and the method further comprises the steps of: inputting the picture information into an identification classification model LRM which has completed training; and carrying out compliance judgment on the model output result.

Further, inputting the picture information into the identification classification model LRM which has completed training, further includes: acquiring self-timer photos uploaded by all staff in a period of time; performing front human body identification extraction on all the self-photographed pictures by using HF-ZF to obtain a plurality of pictures as samples, and performing 360-degree rotation processing on all the sample pictures according to the Y axis of the original pictures to obtain new pictures; taking the original and rotated pictures containing part or all of uniform identification information as positive samples of compliance, and taking the pictures completely without the uniform identification information as negative samples of non-compliance; and (3) reserving parameters of a convolution layer of the inception-v3 model network by utilizing a transfer learning concept, initializing parameters of a full-connection layer, changing model output into 2, taking a training sample which is already classified into the model for training, and obtaining an identification classification model LRM after setting batch, learning rate and epoch value.

Further, the method for judging compliance of the model output result comprises the following steps: setting a judgment threshold value; if the output is a compliance label and the probability value is greater than or equal to the judgment threshold value, judging that the output is compliance; if the output is a compliance label but the probability value is smaller than the judging threshold value, or the output is a non-compliance label, judging that the uniform identification information is not compliance.

A second aspect of the present invention provides a method for detecting information compliance, applied to a client, the method comprising: the method comprises the steps that a client sends a picture to a server, so that the server codes the picture and then carries out compliance judgment on picture information by utilizing a character detection model and an identification information identification model, and if the picture is judged to be non-compliance, the client sends non-compliance information which comprises character non-compliance information or uniform non-compliance information; the client receives and presents the non-compliance information.

Further, the method further comprises: and the user confirms the prompted non-compliance information and sends a new picture to the server.

A third aspect of the present invention provides an apparatus for detecting information compliance, applied to a server, including: the first receiving module is used for receiving the pictures; the coding module is used for performing base64 coding processing on the picture to form picture information; the detection module is used for detecting the character information in the picture information by utilizing the character detection model; the first judging module is used for determining whether the character information is in compliance or not, wherein the character information comprises human body information containing the whole human face characteristics; the extraction module is used for extracting and obtaining a front character region picture if the character information is compliant; the first sending module is used for sending the character non-compliance information to the client if the character non-compliance information is not compliant, so as to remind a user that the character is not compliant, and sending the extracted front character region picture to the encoding module if the character non-compliance information is not compliant; the identification module is used for identifying the character picture information by utilizing the identification information identification model; the second judging module is used for determining whether uniform identification information in the figure picture information is in compliance or not, wherein the uniform identification information comprises an image or a character identification of a unit to which a mark on the uniform belongs; and the second sending module is used for sending the non-compliance information to the client if the uniform is not in compliance, so as to remind the user of the non-compliance of the uniform.

Further, the device further comprises: and the second receiving module is used for responding to the user operation and receiving the new picture.

Further, the device further comprises:

The acquisition module is used for acquiring self-timer photos uploaded by all staff in a period of time; the screening module is used for screening the pictures according to the front characters which can observe the facial features in the pictures; the labeling module is used for labeling the characters on the front face in the picture; the first building module is used for substituting the marked data file into the improved Faster-RCNN to train to obtain a character detection model; the extraction module is used for carrying out front human body identification extraction on all the self-shot pictures; the classification module is used for classifying the extracted front human body original picture and a new picture obtained by rotating the original figure picture into positive and negative samples; and the second building module is used for building the identification classification model based on the transfer learning of the positive and negative sample pictures substituted into the inception-v3 model.

Further, the detection module is configured to: and identifying and extracting the picture information by using the trained character detection model HF-ZF.

Further, the first judging module is configured to: setting a threshold value for the model result softmax probability output, and outputting only frame coordinate points larger than the threshold value; if the length of the output frame coordinate point set is 1, extracting a human body image containing human face features from the original picture according to the frame coordinate points; if the length of the output frame coordinate point set is not equal to 1, judging that the character is not in compliance, wherein the type of the character is not in compliance.

Further, the identification module is configured to: and classifying the picture information by using the trained identification classification model LRM.

Further, the second judging module is configured to: setting a judgment threshold value; if the output is a compliance label and the probability value is greater than or equal to the judgment threshold value, judging that the output is compliance; if the output is a compliance label but the probability value is smaller than the judging threshold value, or the output is a non-compliance label, judging that the uniform identification information is not compliance.

Further, the first establishing module is configured to: based on fast-RCNN, carrying out convolution on a quarter area on a character picture as a face characteristic, carrying out improvement on the face characteristic, carrying out training on an improved model by carrying out the marked data file into a network carrying out convolution on the whole picture, and obtaining a character detection model HF-ZF.

Further, the second establishing module is configured to: and adjusting the proportion of the positive and negative samples to be near 1:1 by using a picture enhancement mode, and carrying the adjusted samples into a inception-v3 model for migration training to obtain an identification classification model.

A fourth aspect of the present invention provides an apparatus for detecting information compliance, applied to a client, including: the sending module is used for sending the picture to the server so that the server can utilize the character detection model and the identification information identification model to identify picture information after encoding the picture, and if the picture is judged to be non-compliant, the sending module is used for sending the non-compliant information to the client; and the receiving module is used for receiving and displaying the non-compliance information.

Further, the device further comprises: and the retransmission module confirms the prompted non-compliance information and transmits the new picture to the server.

A fifth aspect of the present invention provides a server for information compliance detection, comprising: one or more processors, and a storage device. The storage device is used for storing one or more programs. Wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the method of information compliance detection as provided in the first aspect above.

A sixth aspect of the invention provides a computer readable medium having stored thereon executable instructions which, when executed by a processor, cause the processor to perform the method of information compliance detection as provided in the first aspect above.

A seventh aspect of the invention provides a computer program comprising computer executable instructions for implementing the method of information compliance detection provided in the first aspect when executed.

An eighth aspect of the present invention provides a client for information compliance detection, including: one or more processors, and a storage device. The storage device is used for storing one or more programs. Wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the method of information compliance detection as provided in the second aspect above.

A ninth aspect of the invention provides a computer readable medium having stored thereon executable instructions which, when executed by a processor, cause the processor to perform the method of information compliance detection as provided in the second aspect above.

A tenth aspect of the invention provides a computer program comprising computer executable instructions which when executed are adapted to carry out the method of information compliance detection provided by the second aspect.

The information compliance detection method applied to the server has the following beneficial effects:

The technical scheme provided by the invention can be used for detecting and identifying the picture by using the information compliance detection method, and then sending the information which is judged to be non-compliance to the client. The character detection model uses an improved Faster-RCNN algorithm, and the face region features are convolved again and brought into a second convolution network, so that the accuracy of face feature recognition can be increased, the characters recognized based on target detection are all characters with front faces, the influence of other background characters is eliminated, and the operation is different from the traditional character detection model, and the technology creative improvement is achieved. The identification information recognition model is formed by training based on a large number of classified pictures, the original pictures are subjected to Y-axis rotation for 360 degrees in preparation of training samples, the comprehensiveness of the samples is guaranteed, so that the training model is guaranteed to have good generalization performance, and the detection and recognition modes can monitor and judge whether staff wear uniforms according to regulations or not, so that enterprise brand images are maintained, customer satisfaction is improved, site management is more convenient, model expansibility is higher, application value is higher, and compared with related technologies, the method belongs to creative operations.

The information compliance detection method applied to the client has the following beneficial effects:

According to the technical scheme provided by the embodiment of the invention, the picture can be sent to the server, so that the server can utilize the character detection model and the identification information identification model to identify the picture information after encoding the picture, if the picture is judged to be non-compliant, the non-compliant information is sent to the client, the non-compliant information comprises character non-compliant information or uniform non-compliant information, the client prompts a user to carry out corresponding change according to the received non-compliant information, and the client sends the new picture to the server. The method is beneficial to enhancing the supervision of the image and improving the customer satisfaction.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention as claimed.

Drawings

FIG. 1 shows a schematic diagram of an exemplary system architecture of a method of information compliance detection or an apparatus of information compliance detection to which embodiments of the present invention may be applied;

FIG. 2 schematically illustrates a flow chart of a method of information compliance detection applied to a server according to an embodiment of the present invention;

FIG. 3 schematically illustrates a flow chart of a method for information compliance detection applied to a server according to another embodiment of the present invention;

fig. 4 schematically shows a flow chart of a method for information compliance detection applied to a server according to another embodiment of the present invention.

FIG. 5 schematically illustrates a flow chart of a method of information compliance detection applied to a client in accordance with an embodiment of the present invention;

FIG. 6 schematically illustrates a schematic diagram of server and client interactions according to an embodiment of the invention;

FIG. 7 schematically illustrates a block diagram of an apparatus for information compliance detection applied to a server according to an embodiment of the present invention;

FIG. 8 schematically illustrates a block diagram of an apparatus for information compliance detection applied to a server according to another embodiment of the present invention;

FIG. 9 schematically illustrates a block diagram of an apparatus for information compliance detection applied to a server according to another embodiment of the present invention;

FIG. 10 schematically illustrates a block diagram of an apparatus for information compliance detection applied to a client in accordance with an embodiment of the present invention;

FIG. 11 schematically illustrates a block diagram of an apparatus for information compliance detection applied to a client in accordance with another embodiment of the present invention;

FIG. 12 schematically illustrates a block diagram of a computer system of a server according to an embodiment of the invention;

FIG. 13 schematically illustrates a block diagram of a computer system of a client according to an embodiment of the invention;

fig. 14 schematically illustrates a flow chart of a method of information compliance detection according to an embodiment of the present invention.

Detailed Description

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. It should be understood that the description is only illustrative and is not intended to limit the scope of the invention. In addition, in the following description, descriptions of well-known structures and techniques are omitted so as not to unnecessarily obscure the present invention.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The terms "comprises," "comprising," and/or the like, as used herein, specify the presence of stated features, steps, operations, and/or components, but do not preclude the presence or addition of one or more other features, steps, operations, or components.

All terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. It should be noted that the terms used herein should be construed to have meanings consistent with the context of the present specification and should not be construed in an idealized or overly formal manner.

Where a convention analogous to "at least one of A, B and C, etc." is used, in general such a convention should be interpreted in accordance with the meaning of one of skill in the art having generally understood the convention (e.g., "a system having at least one of A, B and C" would include, but not be limited to, systems having a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.). It should also be appreciated by those skilled in the art that virtually any disjunctive word and/or phrase presenting two or more alternative items, whether in the description, claims, or drawings, should be understood to contemplate the possibilities of including one of the items, either of the items, or both. For example, the phrase "a or B" should be understood to include the possibility of "a" or "B", or "a and B".

FIG. 1 is a schematic diagram of an exemplary system architecture of a method of information compliance detection or an apparatus of information compliance detection to which embodiments of the present invention may be applied

As shown in fig. 1, the system architecture 100 may include one or more of terminal devices (101, 102, 103), a network 104, and a server 105. The network 104 is used as a medium for providing a communication link between the terminal devices (101, 102, 103) and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, among others.

It should be understood that the number of terminal devices, networks and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation. For example, the server 105 may be a server cluster formed by a plurality of servers.

A user may interact with the server 105 via the network 104 using the terminal devices (101, 102, 103) to receive or send messages or the like. The terminal devices (101, 102, 103) may be a variety of electronic devices with display screens including, but not limited to, smartphones, tablet computers, portable computers, desktop computers, and the like.

The server 105 may be a server providing various services. For example, the server 105 may obtain a picture from the terminal device 103 (or the terminal device 101 or 102), encode the picture, and then identify the picture information by using the character detection model and the identification information identification model, if it is determined that the picture is not compliant, send non-compliant information to the client, where the non-compliant information includes character non-compliant information or uniform non-compliant information, and the client prompts the user to make a corresponding change according to the received non-compliant information, and sends a new picture to the server, and such detection and identification manner may monitor and determine whether the staff wears uniform according to the rule, so as to maintain the brand image of the enterprise, so that the customer satisfaction is increased, the site management is more convenient, the model expansibility is stronger, and the application value is higher.

In some embodiments, the method for detecting information compliance provided by the embodiments of the present invention is generally performed by the server 105, and accordingly, the device for detecting information compliance is generally disposed in the server 105. In other embodiments, some terminals may have similar functions as servers to perform the method. Therefore, the method for detecting information compliance provided by the embodiment of the invention is not limited to be executed at the server side.

Fig. 2 schematically shows a flow chart of a method for information compliance detection applied to a server according to an embodiment of the invention.

As shown in fig. 2, the method for detecting the information compliance applied to the server includes steps S110 to S130.

In step S110, a picture is received and encoded into picture information, the picture being a self-photograph of the uploaded worker.

In step S120, the person detection model is used to detect the person information in the picture information, and determine whether the person information is in compliance, where the person information includes human body information including the whole face feature, and if the person information is not in compliance, the person information is sent to the client so as to remind the user that the person is not in compliance.

In step S130, if the identification information is not in compliance, the front character region picture is extracted and encoded into picture information, the character picture information is identified by using the identification information identification model, whether the uniform identification information in the character picture information is in compliance is determined, the uniform identification information comprises an image or a character identification of the unit to which the mark on the uniform belongs, and if the uniform identification information is not in compliance, the non-compliance information is sent to the client so as to remind the user of the non-compliance of the uniform.

The method utilizes the character detection model and the identification information identification model to carry out compliance judgment on the picture information, and if the picture information is judged to be non-compliance, the non-compliance information is sent to the client, and the non-compliance information comprises character non-compliance information or uniform non-compliance information so as to remind a user of the picture non-compliance, so that a new picture can be uploaded in time for re-verification. Moreover, the detection mode has stronger expansibility and higher application value, and belongs to innovative operation compared with the related technology.

In some embodiments of the present invention, the uniform may refer to an enterprise tool worn by a person.

In some embodiments of the present invention, the uniform identification information includes an image or text identification of the unit to which the mark on the uniform belongs.

In some embodiments of the invention, the human detection model comprises a human detection model HF-ZF that has undergone a number of pictorial target annotations and completed training.

In some embodiments of the invention, the identification information recognition model comprises an identification classification model LRM that has undergone a number of picture classifications and completed training.

In some embodiments of the present invention, the character detection model HF-ZF is obtained by screening, labeling and training the employee photos uploaded by the client in the past period of time, and after the training is finished, the server may identify whether the front character area containing the face features by using the character detection model HF-ZF when receiving the uploaded image information.

In some embodiments of the present invention, the identification classification model LRM extracts the front character picture through the character detection model HF-ZF based on the employee photo uploaded by the client in the past period, classifies and trains the acquired character picture, and after training is finished, the server may identify whether the uniform identification is not in compliance by using the identification classification model LRM when receiving the character picture information transmitted from the last step.

Fig. 3 schematically shows a flow chart of a method for information compliance detection applied to a server according to another embodiment of the present invention.

As shown in fig. 3, after the above step S130, step S210 is further included.

In step S130, if the identification information is not qualified, the non-compliance information is sent to the client, so as to remind the user of the non-compliance of the uniform.

In step S210, a new picture is received in response to a user operation.

In the method, if the picture is not compliant, the non-compliant information is sent to the client so as to remind the user of the picture being not compliant. When the user sees the non-compliance information displayed by the client, the user can verify the picture in time, if the non-compliance information is true, the server can respond to the user operation to receive a new picture, and if the picture is compliant, the new picture can be used for replacing the existing picture in time.

In some embodiments of the present invention, the uploaded new picture may be a picture that the user re-photographs at the client. For example, when the user views the non-compliance information presented by the client, the picture is checked and the picture is determined to be non-compliance, in which case the user makes a corresponding change and can re-photograph the new picture using the client.

As shown in fig. 4, the method further includes steps S310 to S370 before the step S110.

In step S310, self-timer photographs uploaded by all staff members over a period of time are acquired.

In step S320, the pictures are screened according to the principle that the human body with the complete face in the picture can be observed, so as to obtain a plurality of pictures meeting the condition.

In step S330, a frame of a human body containing a complete face is marked in the picture and saved as a data file in a corresponding format.

In step S340, the marked data file is substituted into the fusion model to be trained to obtain a character detection model HF-ZF.

In step S350, a plurality of pictures obtained by performing frontal human body recognition extraction on all the self-photographed pictures by using HF-ZF are used as samples, and the person pictures are rotated to obtain new pictures.

In step S360, positive and negative sample classification is performed on the extracted character picture and the rotated new picture.

In step S370, based on the transfer learning, training is performed using the classified samples to obtain an identification classification model LRM.

In some embodiments of the present invention, the device source for uploading the photo may be a mobile phone App or a computer client.

In some embodiments of the present invention, as a preliminary sample preparation stage of training of the deep learning target detection algorithm, the number of pictures meeting the conditions obtained by screening should be large enough, at least in the order of thousands, and a total of 2000 pictures meeting the requirements are screened as training samples.

In some embodiments of the present invention, the labeling of the picture sample may be performed using software Labelme in the prior art, or the target labeling of the picture may be implemented by a program written by the user. With respect to specific labeling tools and modes of the picture samples, embodiments of the present invention are not particularly limited.

In some embodiments of the invention, the human detection model is based on the Faster-RCNN model structure. The fast-RCNN (Towards Real-Time Object Detection with Region Proposal Networks) is a model for detecting and identifying related targets of pictures, features are extracted, candidate frames are extracted, and the regression and classification of a boundary frame are integrated in a network, so that the network has higher detection speed and higher detection precision. In the feature extraction stage, the Faster-RCNN adopts a convolutional neural network based on ZF or VGG-16, and a large number of network layers can cause a plurality of low-layer edge feature loss conditions to occur. In the embodiment of the invention, a front character picture needs to be identified and extracted, in order to overcome the defect of the traditional Faster-R-CNN, the influence of the face characteristic is increased, in the stage of the picture characteristic convolution, the marked picture is restored to a fixed size M x N, the values of M and N can be selected, such as 256 x 128,144 x 122 and the like, in the embodiment of the invention, 256 x 128 is selected, an upward quarter area and all picture areas after the fixed size are respectively taken as input and enter a first ZF convolutional neural network and a second ZF convolutional neural network, the second convolutional output characteristic diagram of the first ZF neural network and the fourth convolutional output characteristic diagram of the second ZF neural network are subjected to splice completion characteristic fusion, the third convolutional output characteristic diagram of the first ZF neural network and the fifth output characteristic diagram of the second ZF neural network are subjected to splice completion characteristic fusion as input and are input into a fifth convolutional layer in the second ZF neural network, and the third convolutional output characteristic diagram of the first ZF neural network is subjected to splice completion characteristic fusion as input into a sixth convolutional layer in the second convolutional neural network.

In some embodiments of the invention, a plurality of pictures obtained by carrying out frontal human body recognition extraction on all self-photographed pictures by using HF-ZF are taken as samples, the pictures are taken by a front camera of a mobile phone to form mirror face portraits, and in order to ensure that the samples are balanced and a trained model has stronger generalization capability, all sample pictures are subjected to 360-degree rotation processing according to a Y axis of an original picture to obtain new pictures.

In some embodiments of the present invention, the original and rotated pictures containing some or all of the uniform identification information are taken as positive samples of compliance, and the pictures not containing uniform identification information at all are taken as negative samples of non-compliance. The image classification process can be performed by using some marking software or manually, and the embodiment of the invention is not particularly limited. The ratio of positive and negative samples is maintained at about 1:1, and a relatively small sample size can be amplified by adopting a picture enhancement mode.

In some embodiments of the present invention, a pre-trained picture classification model is selected, and models such as VGG-16, acceptance and the like are commonly used at present. In the embodiment of the invention, a inception-v3 model is selected, a transfer learning concept is utilized, the convolutional layer parameters of a inception-v3 model network are reserved, full-connection layer parameters are initialized, model output is changed to 2, a training sample which is already classified is brought into the model for training, a batch can be set to 100, the learning rate is set to 0.01, and an identification classification model LRM is obtained after 1000 epochs.

Fig. 5 schematically shows a flow chart of a method of information compliance detection applied to a client according to an embodiment of the invention.

As shown in fig. 5, the method for detecting the information compliance of the client includes step S410 and step S420.

In step S410, a picture is sent to a server, so that the server identifies picture information by using a character detection model and an identification information identification model after encoding the picture, and if the picture is determined to be non-compliant, non-compliant information including character non-compliant information or uniform non-compliant information is sent to a client.

In step S420, non-compliance information is received and presented.

According to the technical scheme provided by the embodiment of the invention, the picture can be sent to the server, so that the server can utilize the character detection model and the identification information identification model to identify the picture information after encoding the picture, if the picture is judged to be non-compliant, the non-compliant information is sent to the client, the non-compliant information comprises character non-compliant information or uniform non-compliant information, the client prompts a user to carry out corresponding change according to the received non-compliant information, and the client sends the new picture to the server.

In some embodiments of the present invention, the client may be various electronic devices with photographing functions, and the electronic devices further have functions of receiving and transmitting pictures and text information.

In some embodiments of the present invention, when the user finds that taking the uploaded picture is not acceptable, the method further includes sending the new picture taken by the user to the server. The new picture is a picture taken again by the user at the client including the correct uniform identification. For example, when the user views that the client prompts that the transferred picture is not acceptable, in which case the user makes a corresponding change, the client may be used to re-take the uploaded new picture.

For example, when the client shoots and uploads a picture, the picture is transferred to the gateway system in a consumption queue mode, the gateway system transmits the picture to a coding module deployed in a server in a web service mode under the condition of controlling load balancing, and the coding module performs base64 coding on the picture to obtain picture information and transfers the picture information to the character detection model. And after the character detection model receives the picture information, compliance judgment is carried out on the picture information. And if the person image is judged to be non-compliant, sending person non-compliant information to the client, if the person image is judged to be compliant, extracting the front person region image, coding and then transmitting the front person region image to the identification information recognition model, and after the identification information recognition model receives the person image information, judging the person image to be compliant, and if the person image is judged to be non-compliant, sending uniform non-compliant information to the client.

When the identification information identification model judges compliance, automatically closing the page; when the user sees the non-compliance information prompt presented by the client, a new picture needs to be taken again.

FIG. 6 schematically shows a schematic diagram of server and client interactions according to an embodiment of the invention.

The following is a server and client interaction procedure, which may specifically include S1 to S19, as shown in fig. 6:

S1: the server selects self-shot photos of a plurality of workers.

S2: the server screens the pictures according to the principle that the human body with the complete face in the picture can be observed, and a plurality of pictures meeting the condition are obtained.

S3: the server marks the frame of the human body containing the whole human face in the picture and stores the frame as a data file in a corresponding format.

S4: the server obtains a character detection model HF-ZF by substituting the training sample of the S3 into the fusion model for training.

S5: the server uses a plurality of pictures obtained by carrying out frontal human body recognition extraction on all the self-shot pictures by using HF-ZF as samples, and rotates the character pictures to obtain new pictures.

S6: and the server classifies the extracted character picture and the new rotated picture into positive and negative samples.

S7: the server trains by using the classified samples based on transfer learning to obtain an identification classification model LRM.

S8: the client sends the new picture taken by the person to the server.

S9: and the server performs base64 coding processing on the picture to obtain picture information.

S10: the server identifies the character information in the picture information using the character detection model.

S11: and the server performs compliance judgment on the character information identification result.

S12: if the compliance judgment result is that the character information is not compliant, the server sends the non-compliance information to the client so as to remind the user that the character information is not compliant in the picture.

S13: when a user views the non-compliance information displayed by the client, the picture can be checked in time, and if the non-compliance is true, the client sends the photographed new picture to the server.

S14: and if the compliance judgment result is that the front character region picture is compliance, the server sends the front character region picture to a next module of the server.

S15: and the server performs base64 coding processing on the front character region picture to obtain character picture information.

S16: and the server utilizes the identification information identification model to identify uniform identification information in the picture information.

S17: and the server performs compliance judgment on the identification information identification result.

S18: if the user picture is not compliant, the server sends the non-compliant information to the client so as to remind the user that the uniform identification information is not compliant.

S19: when a user views the non-compliance information displayed by the client, the picture can be checked in time, and if the non-compliance is true, the client sends the photographed new picture to the server.

Fig. 7 schematically shows a block diagram of an apparatus for information compliance detection applied to a server according to an embodiment of the present invention.

As shown in fig. 7, the apparatus 200 for detecting information compliance applied to a server includes a first receiving module 210, an encoding module 220, a detecting module 230, a first judging module 240, an extracting module 250, a first transmitting module 260, an identifying module 270, a second judging module 280, and a second transmitting module 290.

Specifically, the first receiving module 210 is configured to receive a picture.

The encoding module 220 is configured to perform base64 encoding processing on the picture to form picture information.

The detecting module 230 is configured to detect the character information in the picture information by using the character detection model.

The first judging module 240 is configured to determine whether the character information is compliant, where the character information includes human body information including a complete human face feature.

And the extraction module 250 extracts the front character region picture if the character information is compliant.

The first sending module 260 sends the character non-compliance information to the client if the character non-compliance information is not compliant, so as to remind the user of the character non-compliance, and if the character non-compliance information is compliant, obtains the front character region picture and sends the front character region picture to the encoding module.

And an identification module 270 for identifying the character picture information obtained from the encoding module using the identification information identification model.

The second judging module 280 is configured to determine whether uniform identification information in the character picture information is compliant, where the uniform identification information includes an image or a text identifier of a unit to which a mark on the uniform belongs.

The second sending module 290 sends the uniform non-compliance information to the client if the uniform is not compliant, so as to remind the user of uniform non-compliance.

The device 200 for detecting the information compliance applied to the server can accurately extract, identify and judge the picture information by using an information compliance detection method, then send the judging non-compliance result to the client, the client carries out relevant prompt according to the receiving non-compliance information, after the user confirms the non-compliance information, the client sends a new picture to the server, and the detection and identification mode can monitor and judge whether staff wears uniform according to regulations, maintains the brand image of enterprises, improves the customer satisfaction degree, is more convenient in site management, has stronger model expansibility and higher application value, and belongs to innovative operation compared with the related technology.

The apparatus 200 for detecting information compliance applied to a server is used to implement the method for detecting information compliance applied to a server described in the embodiment of fig. 2 according to an embodiment of the present invention.

Fig. 8 schematically shows a block diagram of an apparatus for information compliance detection applied to a server according to another embodiment of the present invention.

As shown in fig. 8, the apparatus 300 for information compliance detection applied to a server further includes a second receiving module 310 in addition to the first receiving module 210, the encoding module 220, the detecting module 230, the first judging module 240, the extracting module 250, the first transmitting module 260, the identifying module 270, the second judging module 280, and the second transmitting module 290 described in fig. 7.

Specifically, the second receiving module 310 is configured to receive a new picture in response to a user operation.

In the apparatus 300 for detecting information compliance applied to a server, if there is non-compliance information in an uploaded picture, the non-compliance information is sent to a client so as to remind a user of the non-compliance of the picture. When the user sees the non-compliance information displayed by the client, the user can verify the picture in time, if the non-compliance information is true, the server can respond to the user operation to receive a new picture, and if the picture is compliant, the new picture can be used for replacing the existing picture in time.

The apparatus 300 for information compliance detection applied to a server is used to implement the method for information compliance detection applied to a server described in the embodiment of fig. 3 according to an embodiment of the present invention.

It is understood that the first receiving module 210, the encoding module 220, the detecting module 230, the first judging module 240, the extracting module 250, the first transmitting module 260, the identifying module 270, the second judging module 280, the second transmitting module 290 and the second receiving module 310 may be combined in one module to be implemented, or any one of the modules may be split into a plurality of modules. Or at least some of the functionality of one or more of the modules may be combined with, and implemented in, at least some of the functionality of other modules. According to embodiments of the invention, at least one of the first receiving module 210, the encoding module 220, the detecting module 230, the first determining module 240, the extracting module 250, the first transmitting module 260, the identifying module 270, the second determining module 280, the second transmitting module 290, and the second receiving module 310 may be implemented at least in part as hardware circuitry, such as a Field Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a system on a chip, a system on a substrate, a system on a package, an Application Specific Integrated Circuit (ASIC), or any other reasonable manner of integrating or packaging the circuitry, or any other hardware or firmware implementation, or an appropriate combination of software, hardware, and firmware implementations. Or at least one of the first receiving module 210, the encoding module 220, the detecting module 230, the first judging module 240, the extracting module 250, the first transmitting module 260, the identifying module 270, the second judging module 280, the second transmitting module 290, and the second receiving module 310 may be at least partially implemented as a computer program module, which may perform the functions of the corresponding module when the program is run by a computer.

Fig. 9 schematically shows a block diagram of an apparatus for information compliance detection applied to a server according to another embodiment of the present invention.

As shown in fig. 9, the apparatus 400 for information compliance detection applied to a server further includes an acquisition module 410, a screening module 420, a labeling module 430, a first establishment module 440, an extraction module 450, a classification module 460, and a second establishment module 470, in addition to the first receiving module 210, the encoding module 220, the detection module 230, the first judgment module 240, the extraction module 250, the first sending module 260, the identification module 270, the second judgment module 280, the second sending module 290, and the second receiving module 310 described in fig. 8.

Specifically, the obtaining module 410 is configured to obtain self-shot photos uploaded by all staff members in a period of time.

And the screening module 420 is configured to screen the picture according to the frontal character that can observe the facial features contained in the picture.

The labeling module 430 is configured to label the front characters in the picture.

The first building block 440 is configured to obtain a person detection model based on the modified fast-RCNN training by substituting the labeled data file.

The extraction module 450 is used for performing frontal human body identification extraction on all the self-shot pictures.

The classification module 460 is configured to classify the extracted front-side human body original picture and a new picture obtained by rotating the original figure picture into positive and negative samples.

The second building module 470 is configured to build the identification classification model based on the positive and negative sample pictures substituted into inception-v3 model for transfer learning.

The device 400 for information compliance detection applied to the server can screen the pictures according to the characters with complete face characteristics in the pictures, obtain the pictures meeting the conditions, mark the characters on the front and bring the characters into model training, and the acquired model can improve the accuracy of character information and uniform identification information identification.

According to an embodiment of the present invention, the apparatus 400 for detecting information compliance applied to a server is used to implement the method for detecting information compliance applied to a server described in the embodiment of fig. 4.

It is understood that the first receiving module 210, the encoding module 220, the detecting module 230, the first judging module 240, the extracting module 250, the first transmitting module 260, the identifying module 270, the second judging module 280, the second transmitting module 290, the second receiving module 310, the obtaining module 410, the screening module 420, the labeling module 430, the first establishing module 440, the extracting module 450, the classifying module 460, and the second establishing module 470 may be combined in one module to be implemented, or any one of them may be split into a plurality of modules. Or at least some of the functionality of one or more of the modules may be combined with, and implemented in, at least some of the functionality of other modules. According to embodiments of the invention, at least one of the first receiving module 210, the encoding module 220, the detecting module 230, the first determining module 240, the extracting module 250, the first transmitting module 260, the identifying module 270, the second determining module 280, the second transmitting module 290, the second receiving module 310, the obtaining module 410, the screening module 420, the labeling module 430, the first establishing module 440, the extracting module 450, the classifying module 460, the second establishing module 470 may be implemented at least in part as hardware circuitry, such as a Field Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a system on chip, a system on a substrate, a system on a package, an Application Specific Integrated Circuit (ASIC), or any other reasonable way of integrating or packaging the circuitry, or in a suitable combination of software, hardware and firmware implementations. Or at least one of the first receiving module 210, the encoding module 220, the detecting module 230, the first judging module 240, the extracting module 250, the first transmitting module 260, the identifying module 270, the second judging module 280, the second transmitting module 290, the second receiving module 310, the obtaining module 410, the screening module 420, the labeling module 430, the first establishing module 440, the extracting module 450, the classifying module 460, and the second establishing module 470 may be at least partially implemented as a computer program module, which may perform the functions of the corresponding modules when the program is run by a computer.

as shown in fig. 10, an apparatus 500 for information compliance detection applied to a client includes a transmitting module 510 and a receiving module 520.

Specifically, the sending module 510 is configured to send a picture to a server, so that the server identifies the picture information by using the person detection model and the identification information identification model after encoding the picture, determines whether there is non-compliance information, and if there is non-compliance, sends the non-compliance information to the client, where the non-compliance information includes person non-compliance information or uniform non-compliance information.

The receiving module 520 is configured to receive and display the non-compliance information.

The apparatus 500 for detecting information compliance applied to a client may send a picture to a server, so that the server encodes the picture and then identifies the picture information using a person detection model and an identification information identification model, and if it is determined that the picture is not compliant, send non-compliance information to the client, where the non-compliance information includes person non-compliance information or uniform non-compliance information, and the client prompts a user to make a corresponding change according to the received non-compliance information, and the client sends a new picture to the server.

According to an embodiment of the present invention, the apparatus 500 for information compliance detection applied to a client is used to implement the method for information compliance detection applied to a server described in the embodiment of fig. 5.

Fig. 11 schematically shows a block diagram of an apparatus for information compliance detection applied to a client according to another embodiment of the present invention.

As shown in fig. 11, the apparatus 600 for information compliance detection applied to a client includes a retransmission module 610 in addition to the transmission module 510 and the reception module 520 described in fig. 10.

Specifically, the retransmission module 610 confirms the presented non-compliance information and sends the new picture to the server.

In the apparatus 600 for detecting information compliance of a client, when a user views non-compliance information displayed by the client, the user checks the picture and determines that the picture is not compliant, in which case the user may take a photograph again using the client to upload a new picture.

It is understood that the transmitting module 510, the receiving module 520, and the retransmission module 610 may be combined in one module to be implemented, or any one of the modules may be split into a plurality of modules. Or at least some of the functionality of one or more of the modules may be combined with, and implemented in, at least some of the functionality of other modules. According to embodiments of the invention, at least one of the transmitting module 510, the receiving module 520, and the retransmission module 610 may be implemented at least in part as hardware circuitry, such as a Field Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a system on a chip, a system on a substrate, a system on a package, an Application Specific Integrated Circuit (ASIC), or any other reasonable way of integrating or packaging circuitry, or in hardware or firmware, or in any suitable combination of software, hardware, and firmware implementations. Or at least one of the transmitting module 510, the receiving module 520, and the retransmitting module 610 may be at least partially implemented as a computer program module, which may perform the functions of the corresponding module when the program is run by a computer.

Fig. 12 schematically shows a block diagram of a computer system of a server according to an embodiment of the invention. The computer system shown in fig. 12 is only an example, and should not be construed as limiting the functionality and scope of use of the embodiments of the invention.

As shown in fig. 12, a computer system 700 of a server for information compliance detection according to an embodiment of the present invention includes a processor 701 which can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) 702 or a program loaded from a storage section 708 into a Random Access Memory (RAM) 703. The processor 701 may include, for example, a general purpose microprocessor (e.g., a CPU), an instruction set processor and/or an associated chipset and/or a special purpose microprocessor (e.g., an Application Specific Integrated Circuit (ASIC)), or the like. The processor 701 may also include on-board memory for caching purposes. The processor 701 may comprise a single processing unit or a plurality of processing units for performing the different actions of the method flows according to embodiments of the invention described with reference to fig. 2-4.

In the RAM 703, various programs and data required for the operation of the system 700 are stored. The processor 701, the ROM 702, and the RAM 703 are connected to each other through a bus 704. The processor 701 performs various steps of the information compliance detection method applied to the server described above with reference to fig. 1 to 2 by executing programs in the ROM 702 and/or the RAM 703. Note that the program may be stored in one or more memories other than the ROM 702 and the RAM 703. The processor 701 may also perform the various steps of the information compliance detection method applied to the server described above with reference to fig. 2-4 by executing a program stored in the one or more memories.

According to an embodiment of the invention, the computer system 700 of the server may further include an input/output (I/O) interface 707, the input/output (I/O) interface 707 also being connected to the bus 704. The system 700 may also include one or more of the following components connected to the I/O interface 705: an input section 706 including a keyboard, a mouse, and the like; an output portion 707 including a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, a speaker, and the like; a storage section 708 including a hard disk or the like; and a communication section 709 including a network interface card such as a LAN card, a modem, or the like. The communication section 709 performs communication processing via a network such as the internet. The drive 710 is also connected to the I/O interface 705 as needed. A removable medium 711 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 710 as necessary, so that a computer program read therefrom is mounted into the storage section 708 as necessary.

According to an embodiment of the present invention, the method described above with reference to the flowcharts may be implemented as a computer software program. For example, embodiments of the present invention include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method shown in the flowcharts. In such an embodiment, the computer program may be downloaded and installed from a network via the communication portion 709, and/or installed from the removable medium 711. The above-described functions defined in the system of the embodiment of the present invention are performed when the computer program is executed by the processor 701. Further, the systems, devices, means, modules, units, etc. described above may be implemented by computer program modules.

The computer readable medium shown in the present invention may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present invention, however, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing. Further, the computer-readable medium may include ROM 702 and/or RAM703 and/or one or more memories other than ROM 702 and RAM703 described above.

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

As another aspect, the present invention also provides a computer-readable medium that may be contained in the apparatus described in the above embodiments; or may be present alone without being fitted into the device. The computer-readable medium carries one or more programs which, when executed by one of the apparatuses, cause the apparatus to perform the information compliance detection method applied to the server according to the embodiment of the present invention. The method comprises the following steps: receiving a picture and encoding the picture into picture information, wherein the picture is an uploaded self-shot picture of a worker; detecting character information in the picture information by using a character detection model, and determining whether the character information is compliant, wherein the character information comprises human body information containing complete human face characteristics, and if the character information is not compliant, sending non-compliant information to a client so as to remind a user that the character is not compliant; if the identification information is not in compliance, sending non-compliance information to a client so as to remind a user of the non-compliance of the uniform; after the non-compliance information is sent to the client, responding to the user operation, and if the staff chooses to re-detect, receiving a new picture.

FIG. 13 schematically illustrates a block diagram of a computer system of a client for information compliance detection, in accordance with an embodiment of the present invention. The computer system shown in fig. 13 is merely an example, and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

As shown in fig. 13, a computer system 800 of a client according to an embodiment of the present invention includes a processor 801 that can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) 802 or a program loaded from a storage section 808 into a Random Access Memory (RAM) 803. The processor 801 may include, for example, a general purpose microprocessor (e.g., a CPU), an instruction set processor and/or an associated chipset and/or special purpose microprocessor (e.g., an Application Specific Integrated Circuit (ASIC)), or the like. The processor 801 may also include on-board memory for caching purposes. The processor 801 may comprise a single processing unit or a plurality of processing units for performing different actions of the method flow according to the embodiment of the invention described with reference to fig. 5.

In the RAM 803, various programs and data required for the operation of the computer system 800 of the client are stored. The processor 801, the ROM 802, and the RAM 803 are connected to each other by a bus 804. The processor 801 performs various steps of the information compliance detection method applied to the client described above with reference to fig. 5 by executing programs in the ROM 802 and/or the RAM 803. Note that the program may be stored in one or more memories other than the ROM 802 and the RAM 803. The processor 801 may also perform the various steps of the information compliance detection method applied to the client described above with reference to fig. 5 by executing a program stored in the one or more memories.

According to an embodiment of the invention, the system 800 may also include an input/output (I/O) interface 807, the input/output (I/O) interface 807 also being connected to the bus 804. The system 800 may also include one or more of the following components connected to the I/O interface 805: an input portion 806 including a keyboard, mouse, etc.; an output portion 807 including a display such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and a speaker; a storage section 808 including a hard disk or the like; and a communication section 809 including a network interface card such as a LAN card, a modem, or the like. The communication section 809 performs communication processing via a network such as the internet. The drive 810 is also connected to the I/O interface 805 as needed. A removable medium 811 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 810 as needed so that a computer program read out therefrom is mounted into the storage section 808 as needed.

The method described above with reference to the flowcharts may be implemented as a computer software program according to embodiments of the present invention. For example, embodiments of the present invention include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method shown in the flowcharts. In such an embodiment, the computer program may be downloaded and installed from a network via the communication section 809, and/or installed from the removable media 811. The above-described functions defined in the system of the embodiment of the present invention are performed when the computer program is executed by the processor 801. Further, the systems, devices, means, modules, units, etc. described above may be implemented by computer program modules.

The computer readable medium shown in the present invention may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present invention, however, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing. Further, the computer-readable medium may include ROM 802 and/or RAM 803 and/or one or more memories other than ROM 802 and RAM 803 described above.

As another aspect, the present invention also provides a computer readable medium for information compliance detection, which may be contained in the apparatus described in the above embodiments; or may be present alone without being fitted into the device. The computer-readable medium carries one or more programs which, when executed by one of the apparatuses, cause the apparatus to perform the information compliance detection method applied to the client according to the embodiment of the present invention. The method comprises the following steps: the method comprises the steps of sending a picture to a server, enabling the server to encode the picture, utilizing a character detection model and an identification information identification model to identify picture information, determining whether non-compliance information exists, and sending the non-compliance information to a client if the non-compliance information exists, wherein the non-compliance information comprises character non-compliance information or uniform non-compliance information; the client receives and displays the non-compliance information; and the staff confirms the non-compliance information displayed by the client and sends a new picture to the server.

The embodiments of the present invention are described above. These examples are for illustrative purposes only and are not intended to limit the scope of the present invention. Although the embodiments are described above separately, this does not mean that the measures in the embodiments cannot be used advantageously in combination. The scope of the invention is defined by the appended claims and equivalents thereof. Various alternatives and modifications can be made by those skilled in the art without departing from the scope of the invention, and such alternatives and modifications are intended to fall within the scope of the invention.

Claims

1. A method of information compliance detection, comprising the steps of:

s1, receiving a picture from a worker client and encoding the picture into picture information, wherein the picture is an uploaded worker self-shot picture;

s2, detecting character information in the picture information by utilizing a character detection model, judging whether the character information is qualified or not, wherein the character information comprises human body information containing the whole face characteristics, and if the character information is not qualified, sending the non-qualified information to a client so as to remind workers that the picture is not qualified, and executing S1 again;

s3, if the uniform is met, extracting a front character region picture, encoding the front character region picture into picture information, identifying character picture information by using an identification information identification model, determining whether uniform identification information in the character picture information is met, wherein the uniform identification information comprises an image or character identification of a unit to which a mark on the uniform belongs, and if the uniform identification information is not met, sending non-uniform information to a client so as to remind a user of the non-uniform;

In the step S3, after sending the non-compliance information to the client, in response to the user operation, if the worker chooses to re-detect, a new picture is received, and S1 is re-executed; or the operator chooses to finish;

wherein:

The character detection model comprises a character detection model HF-ZF which is subjected to a large number of picture target labels and is trained;

The identification information identification model comprises an identification classification model LRM which is used for classifying a large number of pictures and completing training;

Detecting character information in the picture information by using the character detection model, and judging whether the picture information is compliant, wherein the method comprises the following steps: inputting the picture information into a character detection model HF-ZF which has completed training; compliance judgment is carried out on the model output result;

Wherein, input the picture information to the person detection model HF-ZF that has already finished training, still include before: acquiring self-timer photos uploaded by all staff in a period of time; screening the pictures according to the principle that the human body with the complete face in the picture can be observed, so as to obtain a plurality of pictures meeting the condition; marking a frame of a human body containing a whole face in the picture and storing the frame as a data file in a corresponding format; based on a Faster-RCNN model structure, in order to increase the influence of face features, in a picture feature convolution stage, firstly carrying out resolution to a fixed size M x N on a marked picture, respectively taking an upward quarter area and all picture areas after the fixed size as inputs, entering a first ZF convolution neural network and a second ZF convolution neural network, carrying out splicing and completing feature fusion on a second convolution output feature image of the first ZF neural network and a fourth convolution output feature image of the second ZF neural network, carrying out feature capture on a fifth convolution layer which is input into the second convolution neural network, carrying out splicing and completing feature fusion on a third convolution output feature image of the first ZF neural network and a fifth convolution output feature image of the second ZF neural network, and carrying out feature capture on a sixth convolution layer which is input into the second convolution neural network; substituting the marked data file into the adjusted model for training to obtain a character detection model HF-ZF;

The method comprises the following steps of: setting a threshold value for the model result softmax probability output, and outputting only frame coordinate points larger than the threshold value; if the length of the output frame coordinate point set is 1, extracting a human body image containing human face features from the original picture according to the frame coordinate points; if the length of the output frame coordinate point set is not equal to 1, judging that the frame coordinate point set is not compliant, wherein the non-compliant type is character non-compliant;

Identifying the picture information by using an identification information identification model, and determining whether uniform identification information in the picture information is compliant or not comprises the following steps: inputting the picture information into an identification classification model LRM which has completed training;

Compliance judgment is carried out on the model output result;

the method comprises the steps of inputting the picture information into an identification classification model LRM which has completed training, and further comprises the following steps: acquiring self-timer photos uploaded by all staff in a period of time; performing front human body identification extraction on all the self-photographed pictures by using HF-ZF to obtain a plurality of pictures as samples, and performing 360-degree rotation processing on all the sample pictures according to the Y axis of the original pictures to obtain new pictures; taking the original and rotated pictures containing part or all of uniform identification information as positive samples of compliance, and taking the pictures completely without the uniform identification information as negative samples of non-compliance; the method comprises the steps of utilizing a transfer learning concept, reserving parameters of a convolution layer of inception-v3 model network, initializing parameters of a full-connection layer, changing model output into 2, taking a training sample which is already classified into a model for training, and obtaining an identification classification model LRM after batch, learning rate and epoch value are set;

The method comprises the following steps of: setting a judgment threshold value; if the output is a compliance label and the probability value is greater than or equal to the judgment threshold value, judging that the output is compliance; if the output is a compliance label but the probability value is smaller than the judging threshold value, or the output is a non-compliance label, judging that the uniform identification information is not compliance.

2. The method of claim 1, wherein receiving a picture and encoding into picture information comprises: and performing base64 coding processing on the picture to form picture information.

3. The method for detecting information compliance according to claim 1, applied to a client, comprising: the method comprises the steps that a client sends a picture to a server, so that the server codes the picture and then carries out compliance judgment on the picture by utilizing a character detection model and an identification information identification model, and if the picture is judged to be non-compliance, the client sends non-compliance information which comprises character non-compliance information and uniform non-compliance information;

the client receives and presents the non-compliance information.

4. The method for detecting information compliance according to claim 3, applied to a client, further comprising: and the staff confirms the non-compliance information displayed by the client and sends a new picture to the server.

5. The utility model provides a device that information compliance detected, is applied to the customer end, its characterized in that: comprising the following steps:

The sending module is used for sending the picture to the server so that the server can utilize the character detection model and the identification information identification model to identify the picture after encoding the picture, determine whether the non-compliance information exists, and send the non-compliance information to the client if the non-compliance information exists, wherein the non-compliance information comprises character non-compliance information and uniform non-compliance information;

Wherein, input the picture information to the person detection model HF-ZF that has already finished training, still include before: acquiring self-timer photos uploaded by all staff in a period of time; screening the pictures according to the principle that the human body with the complete face in the picture can be observed, so as to obtain a plurality of pictures meeting the condition; marking a frame of a human body containing a whole face in the picture and storing the frame as a data file in a corresponding format; based on a Faster-R-CNN model structure, in order to increase the influence of face features, in a picture feature convolution stage, firstly carrying out resolution to a fixed size M x N on a marked picture, respectively taking an upward quarter region and all picture regions after the fixed size as inputs, entering a first ZF convolution neural network and a second ZF convolution neural network, splicing a second convolution output feature map of the first ZF neural network and a fourth convolution output feature map of the second ZF neural network to finish feature fusion, taking the second convolution output feature map as input to carry out feature capture on a fifth convolution layer in the second convolution neural network, carrying out splicing and finishing feature fusion on a third convolution output feature map of the first ZF neural network and a fifth convolution output feature map of the second ZF neural network, and taking the third convolution output feature map as input to carry out feature capture on a sixth convolution layer in the second convolution neural network; substituting the marked data file into the adjusted model for training to obtain a character detection model HF-ZF;

The method comprises the steps of identifying the picture information by using an identification information identification model, and determining whether uniform identification information in the picture information is compliant or not, wherein the method further comprises the following steps: inputting the picture information into an identification classification model LRM which has completed training; compliance judgment is carried out on the model output result;

The receiving module is used for receiving and displaying the non-compliance information;

and the retransmission module confirms the displayed non-compliance information and transmits the new picture to the server.

6. A client for information compliance detection, comprising:

One or more processors; and

A storage means for storing one or more programs;

Wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the method of any of claims 3-4.

7. An apparatus for detecting information compliance, applied to a server, comprising:

the first receiving module is used for receiving the pictures;

the coding module is used for performing base64 coding processing on the picture to form picture information;

The detection module is used for detecting the character information in the picture by utilizing the character detection model;

the first judging module is used for determining whether the character information is in compliance or not, wherein the character information comprises human body information containing the whole human face characteristics;

The extraction module is used for extracting and obtaining a front character region picture if the character information is compliant;

the first sending module is used for sending the character non-compliance information to the client if the character non-compliance information is not compliant, so as to remind a user of the character non-compliance; if the face character region picture is compliant, sending the extracted face character region picture to an encoding module;

The identification module is used for identifying the character picture information obtained from the encoding module by utilizing the identification information identification model;

The second judging module is used for determining whether uniform identification information in the figure picture information is in compliance or not, wherein the uniform identification information comprises an image or a character identification of a unit to which a mark on the uniform belongs;

the second sending module is used for sending uniform non-compliance information to the client if the uniform is not compliant, so as to remind a user of the uniform non-compliance;

The second receiving module is used for responding to the user operation and receiving the new picture;

The method comprises the steps of inputting the picture information into an identification classification model LRM which has completed training, and further comprises the following steps: acquiring self-timer photos uploaded by all staff in a period of time; performing front human body identification extraction on all the self-photographed pictures by using HF-ZF to obtain a plurality of pictures as samples, and performing 360-degree rotation processing on all the sample pictures according to the Y axis of the original pictures to obtain new pictures; taking the original and rotated pictures containing part or all of uniform identification information as positive samples of compliance, and taking the pictures completely without the uniform identification information as negative samples of non-compliance; and (3) reserving parameters of a convolution layer of the inception-v3 model network by utilizing a transfer learning concept, initializing parameters of a full-connection layer, changing model output into 2, taking a training sample which is already classified into the model for training, and obtaining an identification classification model LRM after setting batch, learning rate and epoch value.

8. The apparatus for detecting information compliance according to claim 7, which is applied to a server, and further comprises:

The acquisition module is used for acquiring self-timer photos uploaded by all staff in a period of time;

The screening module is used for screening the pictures according to the front characters which can observe the facial features in the pictures;

The labeling module is used for labeling the characters on the front face in the picture;

the first building module is used for substituting the marked data file into the improved Faster-RCNN to train to obtain a character detection model;

The extraction module is used for carrying out front human body identification extraction on all the self-shot pictures;

The classification module is used for classifying the extracted front human body original picture and a new picture obtained by rotating the original figure picture into positive and negative samples;

And the second building module is used for building the identification classification model based on the transfer learning of the positive and negative sample pictures substituted into the inception-v3 model.

9. A server for information compliance detection, comprising:

One or more processors; and

A storage means for storing one or more programs;

Wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the method of any of claims 1-2.

10. A computer readable medium having stored thereon executable instructions which when executed by a processor cause the processor to perform the method according to any of claims 1-4.