CN112115952B

CN112115952B - Image classification method, device and medium based on full convolution neural network

Info

Publication number: CN112115952B
Application number: CN202010863562.4A
Authority: CN
Inventors: 袭肖明; 于治楼
Original assignee: Shandong Inspur Scientific Research Institute Co Ltd
Current assignee: Shandong Inspur Scientific Research Institute Co Ltd
Priority date: 2020-08-25
Filing date: 2020-08-25
Publication date: 2022-08-02
Anticipated expiration: 2040-08-25
Also published as: CN112115952A

Abstract

The application discloses a method, equipment and medium for image classification based on a full convolution neural network, which comprises the following steps: according to the full convolution neural network, dividing a target in the image into a plurality of pixels, and determining a target probability value of each pixel; classifying the pixels with the target probability value larger than a preset value into pre-segmented regions, and determining a plurality of key regions; extracting the vector characteristics of each key area, and determining the associated characteristics among a plurality of key areas according to the vector characteristics of each key area; and inputting the associated features among the key areas into a classifier to determine the category of the image. The embodiment of the specification provides a key region learning method based on deep learning, and finally associated features of a plurality of key regions are learned. The learned associated features can more effectively utilize the information of the key regions and can acquire the relevant information of the key regions, thereby being beneficial to further improving the precision of image classification.

Description

Image classification method, device and medium based on full convolution neural network

Technical Field

The present application relates to the field of computer technologies, and in particular, to a method, an apparatus, and a medium for image classification based on a full convolution neural network.

Background

Image classification is an important research field in computer vision, and has been widely applied in the fields of automatic disease diagnosis, identity recognition, target recognition and the like. In recent years, deep learning has made a great progress in image classification. However, the existing target image classification technology has a poor effect in classifying the target images, and cannot classify the target images well.

Disclosure of Invention

In view of this, embodiments of the present application provide an image classification method, device, and medium based on a full convolution neural network, so as to solve the problems that the existing target image classification technology is poor in effect when performing target image classification, and cannot perform good target image classification.

The embodiment of the application adopts the following technical scheme:

the embodiment of the application provides an image classification method based on a full convolution neural network, which comprises the following steps:

according to the full convolution neural network, dividing a target in the image into a plurality of pixels, and determining a target probability value of each pixel;

classifying the pixels with the target probability value larger than a preset value into pre-segmented regions, and determining a plurality of key regions;

extracting the vector characteristics of each key area, and determining the associated characteristics among a plurality of key areas according to the vector characteristics of each key area;

and inputting the associated features among the key areas into a classifier to determine the category of the image.

It should be noted that, the embodiment of the present specification provides a key region learning method based on deep learning, and finally learns the associated features of a plurality of key regions. The learned associated features can more effectively utilize the information of the key regions and can acquire the relevant information of the key regions, thereby being beneficial to further improving the precision of image classification.

Further, the pre-segmented region is a region segmented by a threshold segmentation algorithm.

It should be noted that the basic idea of the threshold segmentation algorithm is to calculate one or more gray threshold values based on the gray features of the image, compare the gray value of each pixel in the image with the threshold values, and finally classify the pixels into appropriate categories according to the comparison results. A plurality of regions can be segmented through a threshold segmentation algorithm, so that a plurality of key regions can be determined subsequently.

Further, before the extracting the vector feature of each key region, the method further includes:

and extracting the minimum bounding rectangle for each key area.

It should be noted that, in the embodiments of the present specification, the key regions are uniformly set to be rectangles, which is convenient for subsequently extracting the vector features of each key region.

Further, the extracting the vector feature of each key region specifically includes:

vector features for each critical area are extracted using the Resnet algorithm.

The ResNet algorithm introduces a residual network structure, the network layer can be deeply processed through the residual network, and the effect of finally extracting the vector characteristics of each key area is better.

Further, the determining the associated features among the plurality of key regions according to the vector features of each key region specifically includes:

determining the correlation between the key areas according to the vector characteristics of each key area;

and determining the correlation characteristics among the plurality of key areas according to the correlation between the convolutional neural network and the key areas.

Further, the determining the correlation between the key regions according to the vector characteristics of each key region specifically includes:

taking each key area as a node, and constructing a sample pair between the nodes;

training a RankSVM algorithm according to the sample pairs among the nodes, and determining the RankSVM algorithm meeting the requirements;

and inputting the feature vector of each key area into the RankSVM algorithm meeting the requirements, and determining the correlation between the nodes, namely determining the correlation between the key areas.

It should be noted that the RankSVM algorithm can convert the ordering problem between key regions into the classification problem of images, and then learn and solve using an SVM classification model.

Further, the determining the correlation characteristics between a plurality of key areas according to the correlation between the convolutional neural network and the key areas specifically includes:

and according to the correlation among the key areas, constructing a relation graph among the key areas, and performing deep learning on the relation graph among the key areas by using a convolutional neural network to obtain the correlation characteristics among the key areas.

Further, the classifier is an SVM classifier.

It should be noted that the SVM classifier is a linear classifier that maximizes the interval in the feature space, and is a binary classification model. Firstly, a linear classifier refers to a linear function; secondly, the maximum interval unfair principle; furthermore, it solves the binary classification problem (into two classes); and the feature space indicates that the object of which the learning classification is performed is the feature data of the sample. The core concept is as follows: support vector samples may play a critical role in the problem of identification. The support vector is also the sample point closest to the classification hyperplane.

The embodiment of the present application further provides an image classification device based on a full convolution neural network, the device includes:

at least one processor; and the number of the first and second groups,

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:

An embodiment of the present application further provides an image classification medium based on a full convolution neural network, in which computer-executable instructions are stored, and the computer-executable instructions are set to:

The embodiment of the application adopts at least one technical scheme which can achieve the following beneficial effects: the embodiment of the specification provides a key region learning method based on deep learning, and finally associated features of a plurality of key regions are learned. The learned associated features can more effectively utilize the information of the key regions and can acquire the relevant information of the key regions, thereby being beneficial to further improving the precision of image classification.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:

fig. 1 is a schematic flowchart of an image classification method based on a full convolution neural network according to an embodiment of the present disclosure;

fig. 2 is a schematic flowchart of an image classification method based on a full convolution neural network according to a second embodiment of the present disclosure.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be described in detail and completely with reference to the following specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

The technical solutions provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings.

Fig. 1 is a schematic flowchart of an image classification method based on a full convolution neural network provided in an embodiment of the present specification, where the embodiment of the present specification may be implemented by an execution unit of an image classification system, and the following steps may be implemented, where the following steps may be specifically implemented:

step S101, according to the full convolution neural network, a target in the image is divided into a plurality of pixels, and a target probability value of each pixel is determined.

In step S101 of the embodiment of the present specification, the target in the image may be segmented into a plurality of pixels by the full convolution neural network, and the target probability value of each pixel may also be determined by the full convolution neural network, where the larger the target probability value is, the more likely the pixel is the target pixel.

And S102, grouping the pixels with the target probability values larger than the preset value into pre-segmented regions, and determining a plurality of key regions.

In step S102 of the embodiment of the present specification, the preset value may be set to 5, that is, when the target probability value of a pixel is greater than or equal to 5, the pixel may be classified into a pre-segmented region. The pre-segmented region may be a region segmented by a threshold segmentation algorithm.

The basic idea of the threshold segmentation algorithm is to calculate one or more gray threshold values based on the gray features of the image, compare the gray value of each pixel in the image with the threshold values, and finally classify the pixels into appropriate classes according to the comparison results.

Step S103, extracting the vector characteristics of each key area, and determining the associated characteristics among a plurality of key areas according to the vector characteristics of each key area.

In step S103 of the embodiment of the present specification, a Resnet algorithm (residual algorithm) may be used to extract vector features of each key region.

Wherein, the Resnet algorithm makes a reference to the input of each layer and learns to form a residual function. The residual function is easier to optimize, and the network layer number can be greatly deepened.

And step S104, inputting the associated features among the key areas into a classifier to determine the category of the image.

In step S104 of the embodiment of the present specification, the classifier is an SVM classifier. An SVM (support Vector machine) classifier is named as a support Vector machine in Chinese, and an SVM essential model is a linear classifier with maximized intervals in a feature space and is a binary classification model. Firstly, a linear classifier refers to a linear function; secondly, the maximum interval unfair principle; furthermore, it solves the binary classification problem (into two classes); and the feature space indicates that the object of which the learning classification is performed is the feature data of the sample. The core concept is as follows: support vector samples may play a critical role in the problem of identification. The support vector is also the closest sample point to the classification hyperplane (Hyper plane).

The embodiment of the specification provides a key region learning method based on deep learning, and finally associated features of a plurality of key regions are learned. The learned associated features can more effectively utilize the information of the key regions and can acquire the relevant information of the key regions, thereby being beneficial to further improving the precision of image classification.

Corresponding to the first embodiment of the present specification, fig. 2 is a schematic flowchart of an image classification method based on a full convolution neural network provided by the second embodiment of the present specification, and the embodiment of the present specification may be implemented by an execution unit of an image classification system, where the following steps may be implemented, and the specific steps may include:

step S201, according to the full convolution neural network, a target in the image is divided into a plurality of pixels, and a target probability value of each pixel is determined.

In step S201 of the embodiment of the present specification, the target in the image may be segmented into a plurality of pixels by the full convolution neural network, and the target probability value of each pixel may also be determined by the full convolution neural network, where the larger the target probability value is, the higher the possibility that the pixel is the target pixel is.

Step S202, grouping the pixels with the target probability value larger than the preset value into pre-segmented regions, and determining a plurality of key regions.

In step S202 of the embodiment of the present specification, the preset value may be set to 5, that is, when the target probability value of a pixel is greater than or equal to 5, the pixel may be classified into a pre-segmented region. The pre-segmented region may be a region segmented by a threshold segmentation algorithm.

In step S203, for each key region, the minimum circumscribed rectangle is extracted.

In step S203 of the embodiment of the present specification, the key regions are uniformly set to be rectangles, so that the vector features of each key region can be extracted subsequently.

Step S204, extracting the vector characteristics of each key area, and determining the associated characteristics among a plurality of key areas according to the vector characteristics of each key area.

In step S204 of the present specification embodiment, a Resnet algorithm (residual algorithm) may be used to extract vector features of each key region.

Determining the associated features among a plurality of key areas according to the vector features of each key area, which specifically comprises the following steps:

It should be noted that, in the embodiments of the present specification, the correlation between the key regions is determined by the vector feature of each key region, and then the correlation features between the plurality of key regions are determined by the convolutional neural network according to the correlation between the key regions. The correlation between the plurality of key areas may be whether the key areas have correlation therebetween, and the correlation characteristic between the plurality of key areas may be a value of the correlation between the key areas.

The determining the correlation between the key areas according to the vector characteristics of each key area specifically comprises:

For example, for node V, the method of constructing the node edge is as follows: and constructing sample pairs by the node V and the rest nodes, training a RankSVM based on the sample pairs, and determining a RankSVM algorithm meeting the requirements. Assuming that K edges are required for a node, the first K samples of the output of the rank svm algorithm need to be found, i.e., the K samples associated with the node V. And calculating the similarity of the node V and the K samples to obtain the correlation information. Wherein, the edges of the nodes are the correlation among the nodes.

The RankSVM algorithm can convert the ordering problem among key regions into the classification problem of images, and then an SVM classification model is used for learning and solving.

The determining the correlation characteristics between a plurality of key areas according to the correlation between the convolutional neural network and the key areas specifically comprises:

Step S205, inputting the associated features between the plurality of key areas into a classifier to determine the category of the image.

In step S205 of the embodiment of the present specification, the classifier is an SVM classifier. An SVM (support Vector machine) classifier is a linear classifier with maximized interval in feature space and is a binary classification model. Firstly, a linear classifier refers to a linear function; secondly, the maximum interval unfair principle; furthermore, it solves the binary classification problem (into two classes); and the feature space indicates that the object of which the learning classification is performed is the feature data of the sample. The core concept is as follows: support vector samples may play a critical role in the problem of identification. The support vector is also the sample point closest to the classification hyperplane.

It should be noted that, in the prior art, since the utilization of the target key area information and the correlation is omitted, the further improvement of the classification performance is limited. In order to solve the problem, the embodiment of the specification firstly provides a key region learning method based on deep learning, key regions containing important target characteristics are learned, then a resnet algorithm is used for extracting vector characteristics of the key regions, each key region is taken as a node, a Rank SVM algorithm is used for learning correlation among the nodes, and a relationship graph of the key regions is constructed. And finally, learning the correlation mapping chart by using a convolutional neural network, and learning more efficient correlation characteristics of a plurality of key regions. The learned associated features can more effectively utilize the information of the key regions and can acquire the relevant information of the key regions, thereby being beneficial to further improving the precision of image classification.

at least one processor; and the number of the first and second groups,

In the 90 s of the 20 th century, improvements in a technology could clearly distinguish between improvements in hardware (e.g., improvements in circuit structures such as diodes, transistors, switches, etc.) and improvements in software (improvements in process flow). However, as technology advances, many of today's process flow improvements have been seen as direct improvements in hardware circuit architecture. Designers almost always obtain the corresponding hardware circuit structure by programming an improved method flow into the hardware circuit. Thus, it cannot be said that an improvement in the process flow cannot be realized by hardware physical modules. For example, a Programmable Logic Device (PLD), such as a Field Programmable Gate Array (FPGA), is an integrated circuit whose Logic functions are determined by programming the Device by a user. A digital system is "integrated" on a PLD by the designer's own programming without requiring the chip manufacturer to design and fabricate application-specific integrated circuit chips. Furthermore, nowadays, instead of manually manufacturing an Integrated Circuit chip, such Programming is often implemented by "logic compiler" software, which is similar to the software compiler used in program development, but the original code before compiling is also written in a specific Programming Language, which is called Hardware Description Language (HDL), and the HDL is not only one kind but many kinds, such as abel (advanced boot Expression Language), ahdl (alternate Language Description Language), communication, CUPL (computer universal Programming Language), HDCal (Java Hardware Description Language), langa, Lola, mylar, HDL, PALASM, rhydl (runtime Description Language), vhjhdul (Hardware Description Language), and vhygl-Language, which are currently used commonly. It will also be apparent to those skilled in the art that hardware circuitry that implements the logical method flows can be readily obtained by merely slightly programming the method flows into an integrated circuit using the hardware description languages described above.

The controller may be implemented in any suitable manner, for example, the controller may take the form of, for example, a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, an Application Specific Integrated Circuit (ASIC), a programmable logic controller, and an embedded microcontroller, examples of which include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, the memory controller may also be implemented as part of the control logic for the memory. Those skilled in the art will also appreciate that, in addition to implementing the controller as pure computer readable program code, the same functionality can be implemented by logically programming method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Such a controller may thus be considered a hardware component, and the means included therein for performing the various functions may also be considered as a structure within the hardware component. Or even means for performing the functions may be regarded as being both a software module for performing the method and a structure within a hardware component.

The systems, devices, modules or units illustrated in the above embodiments may be implemented by a computer chip or an entity, or by a product with certain functions. One typical implementation device is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

For convenience of description, the above devices are described as being divided into various units by function, and are described separately. Of course, the functionality of the units may be implemented in one or more software and/or hardware when implementing the present application.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium which can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.

The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.

The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. An image classification method based on a full convolution neural network, which is characterized by comprising the following steps:

extracting a minimum circumscribed rectangle for each key area;

extracting the vector characteristics of each key area, and determining the associated characteristics among a plurality of key areas according to the vector characteristics of each key area; determining the associated features among the plurality of key areas according to the vector features of each key area, wherein the determining specifically comprises:

determining the correlation characteristics among a plurality of key areas according to the correlation between the convolutional neural network and the key areas; the determining the correlation characteristics between a plurality of key areas according to the correlation between the convolutional neural network and the key areas specifically comprises:

according to the correlation among the key areas, constructing a relation graph among the key areas, and performing deep learning on the relation graph among the key areas by using a convolutional neural network to obtain correlation characteristics among the key areas;

2. The full convolutional neural network-based image classification method as claimed in claim 1, wherein the pre-segmented region is a region segmented by a threshold segmentation algorithm.

3. The image classification method based on the full convolution neural network according to claim 1, wherein the extracting the vector feature of each key region specifically includes:

4. The image classification method based on the full convolution neural network according to claim 1, wherein the determining the correlation between the key regions according to the vector features of each key region specifically includes:

5. The full convolution neural network-based image classification method according to claim 1, wherein the classifier is an SVM classifier.

6. An image classification device based on a full convolution neural network, the device comprising:

at least one processor; and the number of the first and second groups,

classifying the pixels with the target probability value larger than a preset value into pre-segmented regions, and determining a plurality of key regions; extracting a minimum circumscribed rectangle for each key area;

determining correlation characteristics among a plurality of key areas according to the correlation between the convolutional neural network and the key areas; the determining the correlation characteristics between a plurality of key areas according to the correlation between the convolutional neural network and the key areas specifically comprises:

7. An image classification medium based on a full convolutional neural network, having stored thereon computer-executable instructions configured to:

extracting a minimum circumscribed rectangle for each key area;

determining the correlation characteristics among a plurality of key areas according to the correlation between the convolutional neural network and the key areas; the determining the correlation characteristics among a plurality of key areas according to the correlation between the convolutional neural network and the key areas specifically comprises: