CN111898615A - Feature extraction method, device, equipment and medium of object detection model - Google Patents
Feature extraction method, device, equipment and medium of object detection model Download PDFInfo
- Publication number
- CN111898615A CN111898615A CN202010550323.3A CN202010550323A CN111898615A CN 111898615 A CN111898615 A CN 111898615A CN 202010550323 A CN202010550323 A CN 202010550323A CN 111898615 A CN111898615 A CN 111898615A
- Authority
- CN
- China
- Prior art keywords
- network
- feature
- characteristic
- pyramid network
- object detection
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 151
- 238000000605 extraction Methods 0.000 title claims abstract description 44
- 238000000034 method Methods 0.000 claims abstract description 41
- 230000009466 transformation Effects 0.000 claims description 27
- 238000000926 separation method Methods 0.000 claims description 19
- 230000004927 fusion Effects 0.000 claims description 14
- 239000000126 substance Substances 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 6
- 238000010586 diagram Methods 0.000 description 13
- 230000006870 function Effects 0.000 description 9
- 230000006872 improvement Effects 0.000 description 9
- 238000004590 computer program Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 230000005540 biological transmission Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 230000000295 complement effect Effects 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 229920001296 polysiloxane Polymers 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a method, a device, equipment and a medium for extracting characteristics of an object detection model, which comprise the following steps: inputting an image to be detected to a pre-established feature detection model, wherein the feature detection model comprises a deep convolution network, a first feature pyramid network, a second feature pyramid network and an object detection network, the first feature pyramid network constructs a network structure according to the deep convolution network, and the first feature pyramid network and the pre-obtained second feature pyramid network are fused to obtain the object detection network; and extracting the characteristics corresponding to the image to be detected according to the object detection network of the characteristic extraction model. According to the method and the device, the image to be detected is input into the feature detection model, and the first feature pyramid network and the second feature pyramid network in the feature detection model are fused into the object detection network, so that the effect of feature extraction of the feature extraction network can be improved.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a method, an apparatus, a device, and a medium for extracting features of an object detection model.
Background
Deep learning object detection model frameworks, such as SSD, Faster R-CNN, YOLO, etc., can be divided into feature extraction networks and detection networks. The feature extraction network learns the originally input image to obtain a series of abstract features. And the detection network performs position regression and class classification of the detection object on the features of different levels.
In the existing deep learning object detection model, the effect of characteristic extraction network characteristic extraction is poor, and information loss is easy to occur.
Disclosure of Invention
In view of this, embodiments of the present application provide a method, an apparatus, a device, and a medium for extracting features of an object detection model, which are used to solve the problem that the effect of extracting features of an existing feature extraction network is not good.
The embodiment of the application adopts the following technical scheme:
the embodiment of the application provides a feature extraction method of an object detection model, which comprises the following steps:
inputting an image to be detected to a pre-established feature detection model, wherein the feature detection model comprises a deep convolution network, a first feature pyramid network, a second feature pyramid network and an object detection network, the first feature pyramid network constructs a network structure according to the deep convolution network, and the first feature pyramid network and the pre-obtained second feature pyramid network are fused to obtain the object detection network;
and extracting the characteristics corresponding to the image to be detected according to the object detection network of the characteristic extraction model.
It should be noted that, in the embodiment of the present application, the image to be detected is input into the feature detection model, and the first feature pyramid network and the second feature pyramid network in the feature detection model are fused into the object detection network, so that the effect of feature extraction of the feature extraction network can be improved. The first characteristic pyramid network and the second characteristic pyramid network can complement and fuse with each other, and the transmission and expression capacity of the image characteristics in the characteristic detection model is enhanced.
Further, before the image to be detected is input to the pre-established feature detection model, the method further includes:
and taking the characteristic layer of the deep convolutional network as input, and constructing the first characteristic pyramid network according to a preset mode.
It should be noted that the first feature pyramid network is constructed according to a preset manner by taking a feature layer of a deep convolutional network as an input, where the deep convolutional network may be an existing network structure.
Further, the constructing the first feature pyramid network according to a preset mode by using the feature layer of the deep convolutional network as an input specifically includes:
taking a feature layer of the deep convolutional network as an input;
carrying out separation convolution transformation on the features of the last layer of the deep convolution network, and outputting the first layer of the first feature pyramid network;
performing convolution transformation on the characteristics of the first layer of the first characteristic pyramid network, and outputting a first characteristic layer;
carrying out separation convolution transformation on the features of the last but one layer of the deep convolutional network, and outputting a second feature layer;
adding the first characteristic layer and the second characteristic layer to obtain a second layer of the first characteristic pyramid network;
and for the rest layers of the first characteristic pyramid network, performing convolution transformation on the characteristics of the upper layer of the first characteristic pyramid network, outputting a third characteristic layer, performing separation convolution transformation on the characteristics of the corresponding layer of the deep convolution network, outputting a fourth characteristic layer, and adding the third characteristic layer and the fourth characteristic layer to obtain the rest layers of the first characteristic pyramid network.
It should be noted that, the above discloses a specific method for constructing the first feature pyramid network, and the first feature pyramid network can be better constructed.
Further, at least three layers are arranged on the layer number of the first characteristic pyramid network.
Furthermore, the first feature pyramid network has a structure that multiple feature layers are connected from bottom to top; the second feature pyramid network is structured by connecting a plurality of feature layers from top to bottom.
It should be noted that the first feature pyramid network in the bottom-up structure and the second feature pyramid network in the top-down structure may complement and merge with each other, so as to enhance the transmission and expression capability of the image features in the feature detection model.
Further, after the feature layer of the deep convolutional network is used as an input and the first feature pyramid network is constructed according to a preset mode, the method further includes:
and inverting the structure of the first characteristic pyramid network or the second characteristic pyramid network, and then performing characteristic fusion on the first characteristic pyramid network and the second characteristic pyramid network to obtain the object detection network.
It should be noted that, the method for constructing the object detection network disclosed above can better construct the object detection network, so that the object detection network has better ability to identify and detect objects.
Further, the performing feature fusion on the first feature pyramid network and the second feature pyramid network specifically includes:
performing separation convolution transformation on the feature layer in the first feature pyramid network and the corresponding feature layer in the second feature pyramid network to obtain a fifth feature layer corresponding to the first feature pyramid network and a sixth feature layer corresponding to the second feature pyramid network, so that the number of feature channels of the transformed first feature pyramid network is the same as that of feature channels of the second feature pyramid network;
and adding the fifth characteristic layer and the sixth characteristic layer, and performing separation convolution transformation to obtain a seventh characteristic layer so as to complete the characteristic fusion of the first characteristic pyramid network and the second characteristic pyramid network.
It should be noted that, the above-mentioned discloses a specific method for constructing an object detection network, which can better construct the object detection network, so that the object detection network has better ability to detect an object.
The embodiment of the present application further provides a feature extraction device of an object detection model, the device includes:
the detection unit is used for inputting an image to be detected to a pre-established feature detection model, wherein the feature detection model comprises a deep convolution network, a first feature pyramid network, a second feature pyramid network and an object detection network, the first feature pyramid network constructs a network structure according to the deep convolution network, and the first feature pyramid network and the pre-obtained second feature pyramid network are fused to obtain the object detection network;
and the characteristic unit is used for extracting the characteristics corresponding to the image to be detected according to the object detection network of the characteristic extraction model.
The embodiment of the present application further provides a feature extraction device of an object detection model, 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:
inputting an image to be detected to a pre-established feature detection model, wherein the feature detection model comprises a deep convolution network, a first feature pyramid network, a second feature pyramid network and an object detection network, the first feature pyramid network constructs a network structure according to the deep convolution network, and the first feature pyramid network and the pre-obtained second feature pyramid network are fused to obtain the object detection network;
and extracting the characteristics corresponding to the image to be detected according to the object detection network of the characteristic extraction model.
An embodiment of the present application further provides a feature extraction medium for an object detection model, in which computer-executable instructions are stored, where the computer-executable instructions are set to:
inputting an image to be detected to a pre-established feature detection model, wherein the feature detection model comprises a deep convolution network, a first feature pyramid network, a second feature pyramid network and an object detection network, the first feature pyramid network constructs a network structure according to the deep convolution network, and the first feature pyramid network and the pre-obtained second feature pyramid network are fused to obtain the object detection network;
and extracting the characteristics corresponding to the image to be detected according to the object detection network of the characteristic extraction model.
The embodiment of the application adopts at least one technical scheme which can achieve the following beneficial effects: according to the method and the device, the image to be detected is input into the feature detection model, and the first feature pyramid network and the second feature pyramid network in the feature detection model are fused into the object detection network, so that the effect of feature extraction of the feature extraction network can be improved.
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 a feature extraction method of an object detection model according to a first embodiment of the present disclosure;
fig. 2 is a schematic flowchart of a feature extraction method for an object detection model provided in the second embodiment of the present specification;
fig. 3 is a schematic structural diagram of an object detection system provided in the second embodiment of the present disclosure;
fig. 4 is a schematic structural diagram of a feature extraction device of an object detection model according to a third 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 flow chart of a feature extraction method of an object detection model provided in an embodiment of the present specification, where the embodiment of the present specification may be implemented by an object detection system, and the method specifically includes:
step S101, inputting an image to be detected into a pre-established characteristic detection model by the object detection system.
In step S101 in the embodiment of the present description, the feature detection model includes a deep convolutional network, a first feature pyramid network, a second feature pyramid network, and an object detection network, where the first feature pyramid network constructs a network structure according to the deep convolutional network, and the first feature pyramid network and the second feature pyramid network obtained in advance are fused to obtain the object detection network.
In step S101 in the embodiment of this specification, the feature pyramid network may improve the learning ability of the model for the image features, and may extract features of multi-scale and multi-level abstractions.
The structure of the second feature pyramid network can be a top-down feature pyramid network established by a classical convolution model, and the structure of the first feature pyramid network can be a bottom-up feature pyramid network. The first characteristic pyramid network with the bottom-up structure and the second characteristic pyramid network with the top-down structure can complement and fuse with each other, and the transmission and expression capacity of image characteristics in the characteristic detection model is enhanced.
And S102, the object detection system extracts the corresponding features of the image to be detected according to the object detection network of the feature extraction model.
According to the method and the device, the image to be detected is input into the feature detection model, and the first feature pyramid network and the second feature pyramid network in the feature detection model are fused into the object detection network, so that the effect of feature extraction of the feature extraction network can be improved.
Corresponding to the first embodiment of the present specification, fig. 2 is a schematic flow chart of a feature extraction method of an object detection model provided in the second embodiment of the present specification, where the embodiment of the present specification may be implemented by an object detection system, and specifically includes:
step S201, the object detection system takes the feature layer of the deep convolutional network as input, and constructs the first feature pyramid network according to a preset manner.
In step S201 of the embodiment of the present specification, the structure of the first feature pyramid network may be a bottom-up feature pyramid network. The feature layers of the deep convolutional network may be selected as input layers of the first feature pyramid network, where the number of the feature layers of the deep convolutional network is the same as the number of the feature layers of the first feature pyramid network. The number of the first feature pyramid network feature layers is at least three. When the depth convolution network of the feature detection model is deeper, five layers or six layers can be selected, and when the depth convolution network of the feature detection model is shallower, three layers or four layers can be selected. The number of the selected feature layers is the features of different levels of the deep convolutional network, and the resolution ratio is reduced as the level where the feature layers are located is deepened.
In step S201 in the embodiment of this specification, the constructing the first feature pyramid network according to a preset manner with the feature layer of the deep convolutional network as an input includes:
taking a feature layer of the deep convolutional network as an input;
performing separation convolution transformation on the features of the last layer of the deep convolution network, and outputting the first layer of the first feature pyramid network, wherein a separation convolution kernel can be set to be 3x3 in size, and a point-by-point convolution layer is set to be 1x1 in size;
performing convolution transformation on the characteristics of the first layer of the first characteristic pyramid network, and outputting a first characteristic layer;
carrying out separation convolution transformation on the features of the last but one layer of the deep convolutional network, and outputting a second feature layer;
adding the first characteristic layer and the second characteristic layer to obtain a second layer of the first characteristic pyramid network;
and for the rest layers of the first characteristic pyramid network, performing convolution transformation on the characteristics of the upper layer of the first characteristic pyramid network, outputting a third characteristic layer, performing separation convolution transformation on the characteristics of the corresponding layer of the deep convolution network, outputting a fourth characteristic layer, and adding the third characteristic layer and the fourth characteristic layer to obtain the rest layers of the first characteristic pyramid network.
The first characteristic pyramid network can be formed by connecting a plurality of characteristic layers from bottom to top; the second pyramid network of features may be structured such that multiple layers of features are connected from top to bottom.
Referring to fig. 3, which is a schematic structural diagram of the object detection system, a typical deep convolutional neural network is shown on the left side, and the image features (represented by rectangular blocks in the figure) are learned by inputting images from the bottom, receiving inputs of the previous layer in each layer, performing operations such as convolution and the like (represented by line segments with arrows in the figure). The figure shows a bottom-up hierarchy, and as the network layer deepens, the feature resolution is reduced and the number of feature channels is increased. In the deep convolutional neural network, the feature layers from bottom to top are fm in sequenceA、fmB、fmC. At fmA、fmB、fmCAs input, a bottom-up first feature pyramid network is constructed. The method specifically comprises the following steps:
for feature fmA(last layer of deep convolutional network) performing separation convolutional transformation, setting the deep convolutional kernel of the separation convolutional transformation to be 3x3, setting the point-by-point convolution to be 1x1, and outputting the first layer of the first characteristic pyramid network, which can be recorded as bufmA。
For characteristic bufmAConvolution transformation is carried out, and a transformed feature layer is output and is marked as bufm'AFor feature layer fmB(second to last layer of the deep convolutional network) is subjected to separation convolutional transformation, wherein a deep convolutional kernel is set to be 3x3 in size, point-by-point convolution is set to be 1x1 in size, and a transformed feature layer is output and is recorded as fm'B。
Mixing feature layer bufm'AAnd a feature layer fm'BAdding to obtain a second feature layer of the first feature pyramid network, and marking as bufmB。
Similarly, according to the feature layer bufmBAnd fmCObtaining a third characteristic layer of the first characteristic pyramid network, and marking as tdfmC。
Thus, a first feature pyramid network with a bottom-up structure is obtained.
Step S202, the object detection system inverts the structure of the first characteristic pyramid network or the second characteristic pyramid network acquired in advance, and then performs characteristic fusion on the first characteristic pyramid network and the second characteristic pyramid network to obtain the object detection network.
In step S202 in the embodiment of this specification, performing feature fusion on the first feature pyramid network and the second feature pyramid network specifically includes:
performing separation convolution transformation on the feature layer in the first feature pyramid network and the corresponding feature layer in the second feature pyramid network to obtain a fifth feature layer corresponding to the first feature pyramid network and a sixth feature layer corresponding to the second feature pyramid network, so that the number of feature channels of the transformed first feature pyramid network is the same as that of feature channels of the second feature pyramid network;
and adding the fifth characteristic layer and the sixth characteristic layer, and performing separation convolution transformation to obtain a seventh characteristic layer so as to complete the characteristic fusion of the first characteristic pyramid network and the second characteristic pyramid network.
Referring to fig. 3, a schematic diagram of the object detection network obtained by performing feature fusion on the first feature pyramid network and the second feature pyramid network is shown. The third column (the first column on the far left) is a typical feature pyramid network (the second feature pyramid network), which is an existing network structure, and the feature layer is learned from top to bottom. The feature layers of the second feature pyramid network are tdfm respectivelyA、tdfmB、tdfmC. The feature layers of the first feature pyramid network are bufm respectivelyA、bufmB、bufmC. And inverting any one of the two feature pyramid networks, and then performing feature fusion on the equivalent feature layers. For example, for bufmAAnd tdfmCPerforming feature fusion, comprising the following steps:
respectively to bufmAAnd tdfmCSeparating convolution transformation is carried out to make the number of the transformed characteristic channels consistent, then the two groups of characteristic layers are added to obtain a new characteristic layer, and one-time separating convolution transformation is continuously applied to obtain a fused characteristic layer which is marked as fmI。
Same for bufmBAnd tdfmBPerforming feature fusion to obtain fmIITo bufmCAnd tdfmAPerforming feature fusion to obtain fmIII。
Fused feature layer fmI、fmII、fmIIIThe object detection network can be used for a characteristic detection model, and the characteristics are input into the detection model to carry out regression and class classification of the position of a detection object.
It should be noted that the number of the feature layers of the deep convolutional network, the first feature pyramid network, the second feature pyramid network, and the object detection network is the same. The first characteristic pyramid network with the bottom-up structure and the second characteristic pyramid network with the top-down structure can complement and fuse with each other, and the transmission and expression capacity of image characteristics in the characteristic detection model is enhanced.
In step S203, the object detection system constructs the deep convolutional network, the first feature pyramid network, the second feature pyramid network, and the object detection network into a feature detection model.
Step S204, the object detection system inputs the image to be detected to the feature detection model.
In step S205, the object detection system extracts the corresponding features of the image to be detected according to the object detection network of the feature extraction model.
Corresponding to the second embodiment of this specification, fig. 4 is a schematic structural diagram of a feature extraction apparatus of an object detection model provided in the third embodiment of this specification, where the apparatus includes: detection unit 1, feature unit 2.
The detection unit 1 is configured to input an image to be detected to a pre-established feature detection model, where the feature detection model includes a deep convolution network, a first feature pyramid network, a second feature pyramid network, and an object detection network, the first feature pyramid network constructs a network structure according to the deep convolution network, and the first feature pyramid network and the pre-obtained second feature pyramid network are fused to obtain the object detection network.
The feature unit 2 is configured to extract features corresponding to the image to be detected according to the object detection network of the feature extraction model.
The embodiment of the present application further provides a feature extraction device of an object detection model, 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:
inputting an image to be detected to a pre-established feature detection model, wherein the feature detection model comprises a deep convolution network, a first feature pyramid network, a second feature pyramid network and an object detection network, the first feature pyramid network constructs a network structure according to the deep convolution network, and the first feature pyramid network and the pre-obtained second feature pyramid network are fused to obtain the object detection network;
and extracting the characteristics corresponding to the image to be detected according to the object detection network of the characteristic extraction model.
An embodiment of the present application further provides a feature extraction medium for an object detection model, in which computer-executable instructions are stored, where the computer-executable instructions are set to:
inputting an image to be detected to a pre-established feature detection model, wherein the feature detection model comprises a deep convolution network, a first feature pyramid network, a second feature pyramid network and an object detection network, the first feature pyramid network constructs a network structure according to the deep convolution network, and the first feature pyramid network and the pre-obtained second feature pyramid network are fused to obtain the object detection network;
and extracting the characteristics corresponding to the image to be detected according to the object detection network of the characteristic extraction model.
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 making an integrated Circuit chip, such Programming is often implemented by "logic compiler" software, which is similar to a software compiler used in program development and writing, but the original code before compiling is also written by a specific Programming Language, which is called Hardware Description Language (HDL), and HDL is not only one but many, such as abel (advanced Boolean Expression Language), ahdl (alternate Language Description Language), traffic, pl (core unified Programming Language), HDCal, JHDL (Java Hardware Description Language), langue, Lola, HDL, laspam, hardsradware (Hardware Description Language), vhjhd (Hardware Description Language), and vhigh-Language, which are currently used in most common. 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 (10)
1. A method for extracting features of an object detection model, the method comprising:
inputting an image to be detected to a pre-established feature detection model, wherein the feature detection model comprises a deep convolution network, a first feature pyramid network, a second feature pyramid network and an object detection network, the first feature pyramid network constructs a network structure according to the deep convolution network, and the first feature pyramid network and the pre-obtained second feature pyramid network are fused to obtain the object detection network;
and extracting the characteristics corresponding to the image to be detected according to the object detection network of the characteristic extraction model.
2. The method for extracting features of an object detection model according to claim 1, wherein before the image to be detected is input to a pre-established feature detection model, the method further comprises:
and taking the characteristic layer of the deep convolutional network as input, and constructing the first characteristic pyramid network according to a preset mode.
3. The method for extracting features of an object detection model according to claim 2, wherein the constructing the first feature pyramid network according to a preset manner with the feature layer of the deep convolutional network as an input specifically includes:
taking a feature layer of the deep convolutional network as an input;
carrying out separation convolution transformation on the features of the last layer of the deep convolution network, and outputting the first layer of the first feature pyramid network;
performing convolution transformation on the characteristics of the first layer of the first characteristic pyramid network, and outputting a first characteristic layer;
carrying out separation convolution transformation on the features of the last but one layer of the deep convolutional network, and outputting a second feature layer;
adding the first characteristic layer and the second characteristic layer to obtain a second layer of the first characteristic pyramid network;
and for the rest layers of the first characteristic pyramid network, performing convolution transformation on the characteristics of the upper layer of the first characteristic pyramid network, outputting a third characteristic layer, performing separation convolution transformation on the characteristics of the corresponding layer of the deep convolution network, outputting a fourth characteristic layer, and adding the third characteristic layer and the fourth characteristic layer to obtain the rest layers of the first characteristic pyramid network.
4. The feature extraction method of the object detection model according to claim 3, wherein the number of layers of the first feature pyramid network is provided with at least three layers.
5. The feature extraction method of the object detection model according to claim 2, wherein the first feature pyramid network has a structure in which a plurality of feature layers are connected from bottom to top; the second feature pyramid network is structured by connecting a plurality of feature layers from top to bottom.
6. The method for extracting features of an object detection model according to claim 5, wherein after the first feature pyramid network is constructed according to a preset manner with the feature layer of the deep convolutional network as an input, the method further comprises:
and inverting the structure of the first characteristic pyramid network or the second characteristic pyramid network, and then performing characteristic fusion on the first characteristic pyramid network and the second characteristic pyramid network to obtain the object detection network.
7. The method for extracting features of an object detection model according to claim 6, wherein the feature fusion of the first feature pyramid network and the second feature pyramid network specifically includes:
performing separation convolution transformation on the feature layer in the first feature pyramid network and the corresponding feature layer in the second feature pyramid network to obtain a fifth feature layer corresponding to the first feature pyramid network and a sixth feature layer corresponding to the second feature pyramid network, so that the number of feature channels of the transformed first feature pyramid network is the same as that of feature channels of the second feature pyramid network;
and adding the fifth characteristic layer and the sixth characteristic layer, and performing separation convolution transformation to obtain a seventh characteristic layer so as to complete the characteristic fusion of the first characteristic pyramid network and the second characteristic pyramid network.
8. An apparatus for extracting features of an object detection model, the apparatus comprising:
the detection unit is used for inputting an image to be detected to a pre-established feature detection model, wherein the feature detection model comprises a deep convolution network, a first feature pyramid network, a second feature pyramid network and an object detection network, the first feature pyramid network constructs a network structure according to the deep convolution network, and the first feature pyramid network and the pre-obtained second feature pyramid network are fused to obtain the object detection network;
and the characteristic unit is used for extracting the characteristics corresponding to the image to be detected according to the object detection network of the characteristic extraction model.
9. An apparatus for extracting features of an object detection model, the apparatus comprising:
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:
inputting an image to be detected to a pre-established feature detection model, wherein the feature detection model comprises a deep convolution network, a first feature pyramid network, a second feature pyramid network and an object detection network, the first feature pyramid network constructs a network structure according to the deep convolution network, and the first feature pyramid network and the pre-obtained second feature pyramid network are fused to obtain the object detection network;
and extracting the characteristics corresponding to the image to be detected according to the object detection network of the characteristic extraction model.
10. A feature extraction medium of an object detection model, storing computer-executable instructions, the computer-executable instructions configured to:
inputting an image to be detected to a pre-established feature detection model, wherein the feature detection model comprises a deep convolution network, a first feature pyramid network, a second feature pyramid network and an object detection network, the first feature pyramid network constructs a network structure according to the deep convolution network, and the first feature pyramid network and the pre-obtained second feature pyramid network are fused to obtain the object detection network;
and extracting the characteristics corresponding to the image to be detected according to the object detection network of the characteristic extraction model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010550323.3A CN111898615A (en) | 2020-06-16 | 2020-06-16 | Feature extraction method, device, equipment and medium of object detection model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010550323.3A CN111898615A (en) | 2020-06-16 | 2020-06-16 | Feature extraction method, device, equipment and medium of object detection model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111898615A true CN111898615A (en) | 2020-11-06 |
Family
ID=73207679
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010550323.3A Pending CN111898615A (en) | 2020-06-16 | 2020-06-16 | Feature extraction method, device, equipment and medium of object detection model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111898615A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113344862A (en) * | 2021-05-20 | 2021-09-03 | 北京百度网讯科技有限公司 | Defect detection method, defect detection device, electronic equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108846446A (en) * | 2018-07-04 | 2018-11-20 | 国家新闻出版广电总局广播科学研究院 | The object detection method of full convolutional network is merged based on multipath dense feature |
| CN109472298A (en) * | 2018-10-19 | 2019-03-15 | 天津大学 | Depth binary feature pyramid for the detection of small scaled target enhances network |
| CN110580699A (en) * | 2019-05-15 | 2019-12-17 | 徐州医科大学 | Pathological image cell nucleus detection method based on improved fast RCNN algorithm |
| CN111178217A (en) * | 2019-12-23 | 2020-05-19 | 上海眼控科技股份有限公司 | Method and equipment for detecting face image |
-
2020
- 2020-06-16 CN CN202010550323.3A patent/CN111898615A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108846446A (en) * | 2018-07-04 | 2018-11-20 | 国家新闻出版广电总局广播科学研究院 | The object detection method of full convolutional network is merged based on multipath dense feature |
| CN109472298A (en) * | 2018-10-19 | 2019-03-15 | 天津大学 | Depth binary feature pyramid for the detection of small scaled target enhances network |
| CN110580699A (en) * | 2019-05-15 | 2019-12-17 | 徐州医科大学 | Pathological image cell nucleus detection method based on improved fast RCNN algorithm |
| CN111178217A (en) * | 2019-12-23 | 2020-05-19 | 上海眼控科技股份有限公司 | Method and equipment for detecting face image |
Non-Patent Citations (2)
| Title |
|---|
| 许必宵: "基于多尺度特征融合与上下文分析的目标检测技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 黄守志: "基于多特征多尺度卷积神经网络的目标检测算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113344862A (en) * | 2021-05-20 | 2021-09-03 | 北京百度网讯科技有限公司 | Defect detection method, defect detection device, electronic equipment and storage medium |
| CN113344862B (en) * | 2021-05-20 | 2024-04-12 | 北京百度网讯科技有限公司 | Defect detection method, device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6793838B2 (en) | Blockchain-based data processing methods and equipment | |
| CN109034183B (en) | Target detection method, device and equipment | |
| CN112036236B (en) | Image detection method, device and medium based on GhostNet | |
| CN107622080B (en) | Data processing method and equipment | |
| CN109214193B (en) | Data encryption and machine learning model training method and device and electronic equipment | |
| CN111258575B (en) | Page layout processing method and device | |
| CN112308113A (en) | Target identification method, device and medium based on semi-supervision | |
| CN107292002B (en) | Method and device for reconstructing digital core | |
| CN111898615A (en) | Feature extraction method, device, equipment and medium of object detection model | |
| CN112949642B (en) | Character generation method and device, storage medium and electronic equipment | |
| CN116186330A (en) | Video deduplication method and device based on multi-mode learning | |
| CN115221523A (en) | Data processing method, device and equipment | |
| CN114359935A (en) | Model training and form recognition method and device | |
| CN107135137B (en) | Network group identification method and device | |
| CN111539962A (en) | Target image classification method, device and medium | |
| CN111899264A (en) | Target image segmentation method, device and medium | |
| CN111596946A (en) | Recommendation method, device and medium for intelligent contracts of block chains | |
| CN111598092A (en) | Method for determining target area in image, method and device for identifying target | |
| CN111652074A (en) | Face recognition method, device, equipment and medium | |
| CN111523539A (en) | Character detection method and device | |
| CN112685553A (en) | Method, device, equipment and medium for searching and replacing online document | |
| CN112115952B (en) | Image classification method, device and medium based on full convolution neural network | |
| CN107391525B (en) | Instruction processing method and device for database | |
| CN111539961A (en) | Target segmentation method, device and equipment | |
| CN111651450B (en) | Block copy method, device, equipment and medium based on block chain |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201106 |