CN111339963A - Human body image scoring method and device, electronic equipment and storage medium - Google Patents
Human body image scoring method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN111339963A CN111339963A CN202010128706.1A CN202010128706A CN111339963A CN 111339963 A CN111339963 A CN 111339963A CN 202010128706 A CN202010128706 A CN 202010128706A CN 111339963 A CN111339963 A CN 111339963A
- Authority
- CN
- China
- Prior art keywords
- human body
- image
- classification
- body image
- scoring
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/103—Static body considered as a whole, e.g. static pedestrian or occupant recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
Abstract
The application discloses a method and a device for scoring human body images, electronic equipment and a storage medium, and relates to an image recognition technology. The specific implementation scheme is as follows: acquiring a target human body image to be processed; obtaining classification probability values of the target human body image under each classification according to the human body sample images under at least two classifications; and calculating the scoring result of the target human body image according to the score values respectively corresponding to the classifications and the classification probability value. According to the method and the device, the classification probability value of the target human body image under each classification and the score value corresponding to each classification are determined, the target image is directly scored, therefore, the fact that the key point of the human face does not need to be detected is achieved, the scoring result of the human body image can be determined only by adopting a simple classification regression mode, a lot of calculated amount is saved, the scoring result is not only about the human face region, but the whole human body image can be concerned, and therefore the accuracy of scoring the human body image is improved.
Description
Technical Field
The application relates to the technical field of internet, in particular to an image recognition technology, and specifically relates to a method and a device for scoring a human body image, an electronic device and a storage medium.
Background
In the prior art, human body images can be scored, but the existing scoring mode only aims at human face regions, the detection of key points of the human face needs to be carried out, and the human face scoring is carried out based on detection results, but due to the influence of technologies such as a beauty filter, a PS (packet switched) and the like, the score of the human face cannot be directly reflected, the scoring result is inaccurate, and the recommendation and sequencing of subsequent human body images can be influenced.
Disclosure of Invention
The embodiment of the application provides a scoring method and device for human body images, electronic equipment and a storage medium, and aims to solve the technical problem that existing human body images are inaccurate in scoring of attractiveness.
In a first aspect, an embodiment of the present application provides a method for scoring a human body image, including:
acquiring a target human body image to be processed;
obtaining classification probability values of the target human body image under each classification according to the human body sample images under at least two classifications;
and calculating the scoring result of the target human body image according to the score value respectively corresponding to each classification and the classification probability value.
One embodiment in the above application has the following advantages or benefits: the target image is directly scored by determining the classification probability value of the target human body image under each classification and the score value corresponding to each classification, so that the method can determine the scoring result of the human body image only by adopting a simple classification regression mode without detecting key points of the human face, saves a lot of calculated amount, and the scoring result is not only related to the human face region but also can focus on the whole human body image, thereby improving the accuracy of scoring the human body image.
Optionally, obtaining classification probability values of the target human body image under each classification according to human body sample images under at least two classifications includes:
inputting the target human body image into a pre-trained image classification model, and acquiring classification probability values of the target human body image under each classification;
and the image classification model is obtained by training the human body sample images under the at least two classifications.
One embodiment in the above application has the following advantages or benefits: through the pre-trained image classification model, the efficiency and the accuracy of calculating the classification probability value of the target human body image can be improved.
Optionally, before acquiring the target human body image to be processed, the method further includes:
obtaining a plurality of human body sample images marked with classifications in advance;
and training a preset machine learning model by using the plurality of human body sample images to obtain the image classification model.
Optionally, the machine learning model is: a residual network model, wherein a convolution block attention module is superposed in the residual network model.
One embodiment in the above application has the following advantages or benefits: a convolution block attention module is superposed in the residual network model, so that the overall beauty of the human body image can be controlled through an attention mechanism, and the human face image is not limited to the human face value.
Optionally, after obtaining a plurality of human body sample images labeled with classifications in advance, the method further includes:
identifying a head and a body region in each of the human sample images;
according to the head and body areas, on the basis of reserving the head area, cutting processing of at least one body cutting proportion is carried out on each human body sample image, so that human image cutting and increasing are carried out on the human body sample images.
One embodiment in the above application has the following advantages or benefits: the human face image is cut on the basis of reserving the head area, and the influence on the accuracy of scoring the human face image due to the fact that the head area is cut by mistake is avoided.
Optionally, after acquiring the target human body image to be processed, the method further includes:
performing edge zero padding expansion on the target human body image to obtain a square image;
and performing image scaling on the square image to adapt to the input requirement of the residual error network model.
One embodiment in the above application has the following advantages or benefits: the method is characterized in that the edge zero padding expansion is carried out on the human body image, and then the expanded image is zoomed to adapt to the input of the model, so that the method can be ensured to be suitable for the images of various sources and various length-width ratios.
Optionally, before calculating the scoring result of the target human body image according to the score values respectively corresponding to the classifications and the classification probability values, the method further includes:
and obtaining score values respectively corresponding to the classifications in a local searching mode.
One embodiment in the above application has the following advantages or benefits: the score values of the classifications are determined in a local search mode, and the accuracy of each classification score value can be guaranteed.
In a second aspect, an embodiment of the present application further provides a human body image scoring device, including:
the image acquisition module is used for acquiring a target human body image to be processed;
the probability value acquisition module is used for acquiring classification probability values of the target human body image under each classification according to human body sample images under at least two classifications;
and the scoring module is used for calculating the scoring result of the target human body image according to the score value respectively corresponding to each classification and the classification probability value.
In a third aspect, an embodiment of the present application further provides an electronic device, including:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a method of scoring a human body image according to any embodiment of the present application.
In a fourth aspect, the present application further provides a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the method for scoring human body images according to any embodiment of the present application.
One embodiment in the above application has the following advantages or benefits: the classification probability value of the target human body image under each classification and the score value corresponding to each classification are determined, and the target image is directly scored, so that the scoring result of the human body image can be determined only by adopting a simple classification regression mode without detecting key points of the human face, a lot of calculated amount is saved, the scoring result is not only related to the human face region, but also the whole human body image can be concerned, and the accuracy of scoring the human body image is improved; a convolution block attention module is superposed in the residual network model, so that the overall beauty of the human body image can be controlled through an attention mechanism, and the human body image is not limited to the face value; the method is characterized in that the edge zero padding expansion is carried out on the human body image, and then the expanded image is zoomed to adapt to the input of the model, so that the method can be ensured to be suitable for the images of various sources and various length-width ratios.
Other effects of the above-described alternative will be described below with reference to specific embodiments.
Drawings
The drawings are included to provide a better understanding of the present solution and are not intended to limit the present application. Wherein:
fig. 1 is a schematic flow chart of a scoring method for human body images according to a first embodiment of the present application;
FIG. 2 is a flow chart of a scoring method for human body images according to a second embodiment of the present application;
fig. 3 is a schematic structural diagram of a human body image scoring apparatus according to a third embodiment of the present application;
fig. 4 is a block diagram of an electronic device for implementing a scoring method for human body images according to an embodiment of the present application.
Detailed Description
The following description of the exemplary embodiments of the present application, taken in conjunction with the accompanying drawings, includes various details of the embodiments of the application for the understanding of the same, which are to be considered exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present application. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
Fig. 1 is a schematic flow chart of a human body image scoring method according to a first embodiment of the present application, which is applicable to a case of scoring a human body image. The method can be executed by a human body image scoring device which is implemented by software and/or hardware, and is preferably configured in electronic equipment, such as a server and the like. As shown in fig. 1, the method specifically includes the following steps:
s101, obtaining a target human body image to be processed.
The target human body image at least comprises a head area and a human body area, so that the target human body image is scored according to the whole target human body image, and the scoring accuracy is ensured.
S102, obtaining classification probability values of the target human body image under each classification according to the human body sample images under at least two classifications.
Optionally, the human body sample images under the at least two classifications are used for training an image classification model, and then a classification probability value of the target human body image under each classification is determined based on the image classification model. The process of using the image classification model is as follows:
s1, obtaining a plurality of human body sample images marked with classifications in advance.
The classification and labeling of the human body images are exemplary. The image not containing the human body is marked as-1, and the pictures containing the human body are marked as 0,1,2 and 3 according to the aesthetic condition. Thus, for the subsequently trained image classification model, if a pure landscape image is input, a-1 is output, a portrait is input, and a decimal larger than 0, namely a classification probability value, is output.
Further, in order to ensure the accuracy of the image classification model, after acquiring a plurality of human body sample images labeled with classifications in advance, the method further includes:
identifying a head and a body region in each of the human sample images;
according to the head and body areas, on the basis of reserving the head area, cutting processing of at least one body cutting proportion is carried out on each human body sample image, so that human image cutting and increasing are carried out on the human body sample images.
It should be noted here that, in the data augmentation stage, a common method is random cropping, but for portrait pictures, the head of many pictures is in the edge position and is easily cropped due to the fact that the shooting angle is not fixed. The method has the advantages that the head and the body area are firstly and respectively detected, the head is guaranteed not to be cut off, and the accuracy of the image classification model based on human body sample image training is further guaranteed.
Furthermore, because the human body image comes from the eight flowers including a photo, a live screenshot and the like, the aspect ratio of the image is not fixed, and the position where the human face appears is not fixed. After the augmented data is obtained through the cropping augmentation process, the conventional preprocessing means, namely the method of zooming and then cropping into a square, cannot be adopted, because the operation may crop the face. In order to avoid missing key information, the method adopts an image expansion mode of edge zero filling, firstly expands the image into a square, and then carries out scaling operation. Specifically, performing edge zero padding expansion on the human body sample image to obtain a square image; and performing image scaling on the square image to adapt to the input requirement of the residual error network model. Therefore, the precision of the training sample can be ensured, and the precision of the image classification model can be further ensured.
And S2, training a preset machine learning model by using the plurality of human body sample images to obtain the image classification model.
The machine learning model is a residual network model, and the residual network model is a rescet 50_ vd model as an example, wherein a Convolutional Block Attention Module (CBAM) is superimposed in the residual network model, a channel attention and a spatial attention are integrated in the CBAM, during actual training, a pre-trained rescet 50_ vd model is loaded first, then a front layer is frozen, only a part of a rear layer matched with the CBAM is trained, specifically, the target human body image is input into a pre-trained image classification model, and a classification probability value of the target human body image under each classification is obtained. It should be noted that, by superimposing the convolution block attention module on the residual network model, the attention mechanism can compare the overall beauty of the human body image, and not only the face value, so as to ensure the accuracy of final scoring of the human body image.
S103, calculating a scoring result of the target human body image according to the score values respectively corresponding to the classifications and the classification probability value.
The score value of each classification may be preset, and when the scoring result is calculated subsequently, for example, only the score values of the classifications are multiplied by the respective classification probability values, and are summed to obtain the final scoring result.
It should be noted here that, in the prior art, the face region is mainly scored, the detection of the face key points is required before scoring, and the calculation amount is large, but in the present application, after the classification of the degree of quality of a plurality of training samples is labeled in advance, the image classification model is trained by using the training samples, so that the detection of the face key points is converted into the classification regression problem, and the scoring result is not only about the face region, but also can concern the whole human body image.
In the embodiment of the application, the target image is directly scored by determining the classification probability value of the target human body image under each classification and the score value corresponding to each classification, so that the purpose that the human face key point is not required to be detected is achieved, the scoring result of the human body image can be determined only by adopting a simple classification regression mode, a lot of calculated amount is saved, and the scoring result is obtained based on the human body image integrally, so that the accuracy of scoring the human body image is improved.
Fig. 2 is a schematic flow chart of a human body image scoring method according to a second embodiment of the present application, and the present embodiment is further optimized based on the above embodiments. As shown in fig. 2, the method specifically includes the following steps:
s201, obtaining a target human body image to be processed.
S202, conducting edge zero filling expansion on the target human body image to obtain a square image.
S203, performing image expansion and contraction on the square image to adapt to the input requirement of the residual error network model.
As the target human body image to be processed comes from the Fidelity and the like, including a photo, a live screenshot and the like, the aspect ratio of the image is uncertain, and the position of the face is uncertain. The usual pre-processing means, i.e. scaling before cutting into squares, cannot be used, since such operations may cut out the face. In order to avoid missing key information, the method expands the image into a square by adopting an image expansion mode of edge zero padding through steps S202-S203, and then performs scaling operation. Therefore, the key information of the target human body image is kept, and the processed target human body image is adapted to the input requirement of the residual error network model, namely the input requirement of the image classification model.
S204, inputting the target human body image into a pre-trained image classification model, and acquiring classification probability values of the target human body image under each classification.
S205, obtaining score values corresponding to the classifications respectively in a local searching mode.
In order to ensure that the score values of all the classifications are optimal and further ensure that the final scoring result is accurate, the score values of all the classifications are determined in a local searching mode. For example, after the score values of the first three classifications are fixed, the selectable score value of the fourth classification is set in a traversing manner, the selectable score with the highest accuracy is obtained and is used as the optimal score value of the fourth classification, and so on until the optimal score values of all four classifications are obtained.
And S206, calculating a scoring result of the target human body image according to the score values respectively corresponding to the classifications and the classification probability value.
According to the technical scheme, the image expansion mode of edge zero padding is adopted, the image is expanded into a square shape, then scaling operation is carried out, key information is guaranteed not to be lost, and the processed image is guaranteed to be adaptive to the input requirement of the image classification model. Meanwhile, the optimal score value of each classification is determined in a local search mode, and the accuracy of the final scoring result is further guaranteed.
Fig. 3 is a schematic structural diagram of a human body image scoring device according to a third embodiment of the present application, to which this embodiment is applicable. The device can realize the human body image scoring method in any embodiment of the application. As shown in fig. 3, the apparatus 300 specifically includes:
an image acquisition module 301, configured to acquire a target human body image to be processed;
a probability value obtaining module 302, configured to obtain classification probability values of the target human body image under each classification according to human body sample images under at least two classifications;
and the scoring module 303 is configured to calculate a scoring result of the target human body image according to the score values respectively corresponding to the classifications and the classification probability value.
Optionally, the probability value obtaining module is specifically configured to:
inputting the target human body image into a pre-trained image classification model, and acquiring classification probability values of the target human body image under each classification;
and the image classification model is obtained by training the human body sample images under the at least two classifications.
Optionally, the apparatus further comprises:
the device comprises an acquisition unit, a classification unit and a classification unit, wherein the acquisition unit is used for acquiring a plurality of human body sample images marked with classifications in advance;
and the training unit is used for training a preset machine learning model by using the plurality of human body sample images to obtain the image classification model.
Optionally, the machine learning model is: a residual network model, wherein a convolution block attention module is superposed in the residual network model.
Optionally, the apparatus further comprises:
the identification module is used for respectively identifying a head area and a body area in each human body sample image;
and the cutting module is used for performing cutting processing of at least one body cutting proportion on each human body sample image on the basis of reserving the head area according to the head area and the body area so as to cut and expand the human body sample image.
Optionally, the apparatus further comprises:
the expansion module is used for carrying out edge zero filling expansion on the target human body image so as to obtain a square image;
and the stretching module is used for performing image stretching on the square image so as to adapt to the input requirement of the residual error network model.
Optionally, the apparatus further comprises:
and the score value acquisition module is used for acquiring score values corresponding to the classifications respectively in a local search mode.
The human body image scoring device 300 provided by the embodiment of the application can execute the human body image scoring method provided by any embodiment of the application, and has corresponding functional modules and beneficial effects of the execution method. Reference may be made to the description of any method embodiment of the present application for details not explicitly described in this embodiment.
According to an embodiment of the present application, an electronic device and a readable storage medium are also provided.
As shown in fig. 4, it is a block diagram of an electronic device of a scoring method for human body images according to an embodiment of the present application. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present application that are described and/or claimed herein.
As shown in fig. 4, the electronic apparatus includes: one or more processors 401, memory 402, and interfaces for connecting the various components, including high-speed interfaces and low-speed interfaces. The various components are interconnected using different buses and may be mounted on a common motherboard or in other manners as desired. The processor may process instructions for execution within the electronic device, including instructions stored in or on the memory to display graphical information of a GUI on an external input/output apparatus (such as a display device coupled to the interface). In other embodiments, multiple processors and/or multiple buses may be used, along with multiple memories and multiple memories, as desired. Also, multiple electronic devices may be connected, with each device providing portions of the necessary operations (e.g., as a server array, a group of blade servers, or a multi-processor system). In fig. 4, one processor 401 is taken as an example.
The memory 402, as a non-transitory computer readable storage medium, may be used to store non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules (e.g., the image acquisition module 301, the probability value acquisition module 302, and the scoring module 303 shown in fig. 3) corresponding to the scoring method for human body images in the embodiment of the present application. The processor 401 executes various functional applications of the server and data processing by running non-transitory software programs, instructions and modules stored in the memory 402, that is, implements the scoring method of human body images in the above-described method embodiments.
The memory 402 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to use of an electronic device implementing the scoring method of human body images of the embodiment of the present application, and the like. Further, the memory 402 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory 402 may optionally include a memory remotely disposed with respect to the processor 401, and these remote memories may be connected to an electronic device implementing the scoring method for human body images of the embodiments of the present application through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The electronic device for implementing the method for scoring human body images according to the embodiment of the present application may further include: an input device 403 and an output device 404. The processor 401, the memory 402, the input device 403 and the output device 404 may be connected by a bus or other means, and fig. 4 illustrates an example of a connection by a bus.
The input device 403 may receive input numeric or character information and generate key signal inputs related to user settings and function control of an electronic apparatus implementing the scoring method for human body images of the embodiment of the present application, such as an input device of a touch screen, a keypad, a mouse, a track pad, a touch pad, a pointing stick, one or more mouse buttons, a track ball, a joystick, or the like. The output devices 404 may include a display device, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibrating motors), among others. The display device may include, but is not limited to, a Liquid Crystal Display (LCD), a Light Emitting Diode (LED) display, and a plasma display. In some implementations, the display device can be a touch screen.
Various implementations of the systems and techniques described here can be realized in digital electronic circuitry, integrated circuitry, application specific ASICs (application specific integrated circuits), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
These computer programs (also known as programs, software applications, or code) include machine instructions for a programmable processor, and may be implemented using high-level procedural and/or object-oriented programming languages, and/or assembly/machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and/or device (e.g., magnetic discs, optical disks, memory, Programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or data to a programmable processor.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), the internet, and blockchain networks.
The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.
According to the technical scheme of the embodiment of the application, the classification probability values of the target human body image under each classification and the score value corresponding to each classification are determined, and the target image is directly scored, so that the goal of determining key points of the human face is achieved, the scoring result of the human body image can be determined only by adopting a simple classification regression mode, a lot of calculated amount is saved, the scoring result is not only related to the human face region, but the whole human body image can be concerned, and the accuracy of scoring the human body image is improved; a convolution block attention module is superposed in the residual network model, so that the overall beauty of the human body image can be controlled through an attention mechanism, and the human body image is not limited to the face value; the method is characterized in that the edge zero padding expansion is carried out on the human body image, and then the expanded image is zoomed to adapt to the input of the model, so that the method can be ensured to be suitable for the images of various sources and various length-width ratios.
It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present application may be executed in parallel, sequentially, or in different orders, and the present invention is not limited thereto as long as the desired results of the technical solutions disclosed in the present application can be achieved.
The above-described embodiments should not be construed as limiting the scope of the present application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (16)
1. A scoring method for human body images is characterized by comprising the following steps:
acquiring a target human body image to be processed;
obtaining classification probability values of the target human body image under each classification according to the human body sample images under at least two classifications;
and calculating the scoring result of the target human body image according to the score value respectively corresponding to each classification and the classification probability value.
2. The method of claim 1, wherein obtaining classification probability values of the target human body image under each of at least two classifications according to the human body sample image under each classification comprises:
inputting the target human body image into a pre-trained image classification model, and acquiring classification probability values of the target human body image under each classification;
and the image classification model is obtained by training the human body sample images under the at least two classifications.
3. The method according to claim 2, before acquiring the target human body image to be processed, further comprising:
obtaining a plurality of human body sample images marked with classifications in advance;
and training a preset machine learning model by using the plurality of human body sample images to obtain the image classification model.
4. The method of claim 3, wherein the machine learning model is: a residual network model, wherein a convolution block attention module is superposed in the residual network model.
5. The method of claim 3, further comprising, after acquiring the plurality of human body sample images pre-labeled with the classification:
identifying a head and a body region in each of the human sample images;
according to the head and body areas, on the basis of reserving the head area, cutting processing of at least one body cutting proportion is carried out on each human body sample image, so that human image cutting and increasing are carried out on the human body sample images.
6. The method according to claim 5, further comprising, after acquiring the target human body image to be processed:
performing edge zero padding expansion on the target human body image to obtain a square image;
and performing image scaling on the square image to adapt to the input requirement of the residual error network model.
7. The method according to claim 1, further comprising, before calculating the scoring result of the target human body image according to the score values respectively corresponding to the classifications and the classification probability values:
and obtaining score values respectively corresponding to the classifications in a local searching mode.
8. A scoring device for human body images, comprising:
the image acquisition module is used for acquiring a target human body image to be processed;
the probability value acquisition module is used for acquiring classification probability values of the target human body image under each classification according to human body sample images under at least two classifications;
and the scoring module is used for calculating the scoring result of the target human body image according to the score value respectively corresponding to each classification and the classification probability value.
9. The apparatus of claim 8, wherein the root probability value obtaining module is specifically configured to:
inputting the target human body image into a pre-trained image classification model, and acquiring classification probability values of the target human body image under each classification;
and the image classification model is obtained by training the human body sample images under the at least two classifications.
10. The apparatus of claim 9, further comprising:
the device comprises an acquisition unit, a classification unit and a classification unit, wherein the acquisition unit is used for acquiring a plurality of human body sample images marked with classifications in advance;
and the training unit is used for training a preset machine learning model by using the plurality of human body sample images to obtain the image classification model.
11. The apparatus of claim 10, wherein the machine learning model is: a residual network model, wherein a convolution block attention module is superposed in the residual network model.
12. The apparatus of claim 10, further comprising:
the identification module is used for respectively identifying a head area and a body area in each human body sample image;
and the cutting module is used for performing cutting processing of at least one body cutting proportion on each human body sample image on the basis of reserving the head area according to the head area and the body area so as to cut and expand the human body sample image.
13. The apparatus of claim 12, further comprising:
the expansion module is used for carrying out edge zero filling expansion on the target human body image so as to obtain a square image;
and the stretching module is used for performing image stretching on the square image so as to adapt to the input requirement of the residual error network model.
14. The apparatus of claim 8, further comprising:
and the score value acquisition module is used for acquiring score values corresponding to the classifications respectively in a local search mode.
15. An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of scoring of human images of any one of claims 1-7.
16. A non-transitory computer readable storage medium storing computer instructions for causing a computer to execute the scoring method for human body images according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010128706.1A CN111339963A (en) | 2020-02-28 | 2020-02-28 | Human body image scoring method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010128706.1A CN111339963A (en) | 2020-02-28 | 2020-02-28 | Human body image scoring method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111339963A true CN111339963A (en) | 2020-06-26 |
Family
ID=71187206
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010128706.1A Pending CN111339963A (en) | 2020-02-28 | 2020-02-28 | Human body image scoring method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111339963A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114663671A (en) * | 2022-02-21 | 2022-06-24 | 佳都科技集团股份有限公司 | Target detection method, device, equipment and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106296701A (en) * | 2016-08-15 | 2017-01-04 | 北京小米移动软件有限公司 | Image processing method and device |
| CN107590460A (en) * | 2017-09-12 | 2018-01-16 | 北京达佳互联信息技术有限公司 | Face classification method, apparatus and intelligent terminal |
| CN108446659A (en) * | 2018-03-28 | 2018-08-24 | 百度在线网络技术（北京）有限公司 | Method and apparatus for detecting facial image |
| CN109345522A (en) * | 2018-09-25 | 2019-02-15 | 北京市商汤科技开发有限公司 | A kind of picture quality screening technique and device, equipment and storage medium |
| CN109657539A (en) * | 2018-11-05 | 2019-04-19 | 深圳前海达闼云端智能科技有限公司 | Face value evaluation method and device, readable storage medium and electronic equipment |
| CN110390704A (en) * | 2019-07-11 | 2019-10-29 | 深圳追一科技有限公司 | Image processing method, device, terminal device and storage medium |
| CN110796064A (en) * | 2019-10-25 | 2020-02-14 | 清华大学 | Human muscle image establishing method and device, storage medium and electronic equipment |
-
2020
- 2020-02-28 CN CN202010128706.1A patent/CN111339963A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106296701A (en) * | 2016-08-15 | 2017-01-04 | 北京小米移动软件有限公司 | Image processing method and device |
| CN107590460A (en) * | 2017-09-12 | 2018-01-16 | 北京达佳互联信息技术有限公司 | Face classification method, apparatus and intelligent terminal |
| CN108446659A (en) * | 2018-03-28 | 2018-08-24 | 百度在线网络技术（北京）有限公司 | Method and apparatus for detecting facial image |
| CN109345522A (en) * | 2018-09-25 | 2019-02-15 | 北京市商汤科技开发有限公司 | A kind of picture quality screening technique and device, equipment and storage medium |
| CN109657539A (en) * | 2018-11-05 | 2019-04-19 | 深圳前海达闼云端智能科技有限公司 | Face value evaluation method and device, readable storage medium and electronic equipment |
| CN110390704A (en) * | 2019-07-11 | 2019-10-29 | 深圳追一科技有限公司 | Image processing method, device, terminal device and storage medium |
| CN110796064A (en) * | 2019-10-25 | 2020-02-14 | 清华大学 | Human muscle image establishing method and device, storage medium and electronic equipment |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114663671A (en) * | 2022-02-21 | 2022-06-24 | 佳都科技集团股份有限公司 | Target detection method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110659600B (en) | Object detection method, device and equipment | |
| CN111277912B (en) | Image processing method and device and electronic equipment | |
| CN111860479A (en) | Optical character recognition method, device, electronic equipment and storage medium | |
| CN111401251B (en) | Lane line extraction method, lane line extraction device, electronic equipment and computer readable storage medium | |
| CN112528850A (en) | Human body recognition method, device, equipment and storage medium | |
| CN111753961A (en) | Model training method and device, and prediction method and device | |
| CN111709878A (en) | Face super-resolution implementation method and device, electronic equipment and storage medium | |
| CN111722245A (en) | Positioning method, positioning device and electronic equipment | |
| CN111860167A (en) | Face fusion model acquisition and face fusion method, device and storage medium | |
| CN111783948A (en) | Model training method and device, electronic equipment and storage medium | |
| CN111611903B (en) | Training method, using method, device, equipment and medium of motion recognition model | |
| US20210334602A1 (en) | Method and Apparatus for Recognizing Text Content and Electronic Device | |
| CN111462029A (en) | Visual point cloud and high-precision map fusion method and device and electronic equipment | |
| CN111968203A (en) | Animation driving method, animation driving device, electronic device, and storage medium | |
| CN112241716B (en) | Training sample generation method and device | |
| CN111783760A (en) | Character recognition method and device, electronic equipment and computer readable storage medium | |
| CN111090991A (en) | Scene error correction method and device, electronic equipment and storage medium | |
| CN110532415B (en) | Image search processing method, device, equipment and storage medium | |
| CN111539347A (en) | Method and apparatus for detecting target | |
| CN111191619A (en) | Method, device and equipment for detecting virtual line segment of lane line and readable storage medium | |
| CN112508964B (en) | Image segmentation method, device, electronic equipment and storage medium | |
| CN112508027B (en) | Head model for instance segmentation, instance segmentation model, image segmentation method and device | |
| CN111339963A (en) | Human body image scoring method and device, electronic equipment and storage medium | |
| CN112488126A (en) | Feature map processing method, device, equipment and storage medium | |
| CN111696134A (en) | Target detection method and device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |