CN113139577B - Deep learning image classification method and system based on deformable convolution network - Google Patents
Deep learning image classification method and system based on deformable convolution network Download PDFInfo
- Publication number
- CN113139577B CN113139577B CN202110303551.5A CN202110303551A CN113139577B CN 113139577 B CN113139577 B CN 113139577B CN 202110303551 A CN202110303551 A CN 202110303551A CN 113139577 B CN113139577 B CN 113139577B
- Authority
- CN
- China
- Prior art keywords
- layer
- image
- convolution
- gray
- network
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000013135 deep learning Methods 0.000 title claims abstract description 24
- 238000012549 training Methods 0.000 claims abstract description 37
- 238000012360 testing method Methods 0.000 claims abstract description 16
- 210000005036 nerve Anatomy 0.000 claims description 12
- 238000011176 pooling Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 4
- 238000000605 extraction Methods 0.000 abstract description 5
- 238000004364 calculation method Methods 0.000 abstract description 4
- 241000282994 Cervidae Species 0.000 description 2
- 241000282326 Felis catus Species 0.000 description 2
- 241000009328 Perro Species 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000008092 positive effect Effects 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a deep learning image classification method and a system based on a deformable convolution network, which relate to the field of computer vision image recognition, and the method comprises the following steps: step S1: scaling all images to 90 x 90 size and dividing them into training set and test set; step S2: setting a deformable convolution network structure; step S3: discarding part of convolution kernel parameters of the second layer and the fourth layer of the deformable convolution network structure according to a certain rule; step S4: training the training set data according to the network model; step S5: the trained model performs classification recognition on the test set data (assuming that N classes of images are to be classified). The method of the invention discards part of convolution kernel parameters in the network through the processing of a certain rule, improves the key information extraction capability of the network, reduces network parameters and calculation amount, improves the network classification accuracy and improves the network operation efficiency.
Description
Technical Field
The invention relates to the field of computer vision image recognition, in particular to a deep learning image classification method and system based on a deformable convolution network.
Background
In recent years, deep learning is widely applied in academia and scientific sciences, especially in the image field, and has been greatly progressed in the image classification field at present, so that good effects are achieved. When people observe the environment, the brain usually only focuses on a certain part with special importance, acquires key information and removes insignificant information, but the current mainstream deep learning networks such as VGG16 and ALexNet, googleNet have weaker extraction capability on the key information of the images, and the useful characteristic information of the images is extracted and contains some insignificant information, so that the performance of a network model is affected.
Disclosure of Invention
In view of the foregoing, it is necessary to provide a method and a system for classifying deep learning images based on a deformable convolutional network, which automatically designs a convolutional kernel according to feature information of classified images, discards part of parameters, and reduces network operation calculation and improves network operation efficiency while improving the extraction capability of key feature information of the network.
In order to achieve the above purpose, the present invention provides the following technical solutions:
in one aspect, the invention provides a deep learning image classification method based on a deformable convolutional network, comprising the following steps:
step S1: scaling all images to 90 x 90 size and dividing them into training set and test set;
step S2: setting a deformable convolution network structure;
step S3: discarding part of convolution kernel parameters of the second layer and the fourth layer of the deformable convolution network structure according to a certain rule:
step S4: training the training set data according to the network model;
step S5: and classifying and identifying the test set data by the trained model.
Further, in step S2, the deformable convolutional network structure specifically includes: layer 1 is an input layer, and 90 x 90 pictures are input; the layer 2 is a convolution layer, the number of convolution kernels is 16, and the size of the convolution kernels is 9*9; layer 3 is the maximum pooling layer, with a core size of 2 x 2; the 4 th layer is a convolution layer, the number of convolution kernels is 16, and the size of the convolution kernels is 9*9; layer 5 is the maximum pooling layer, the core size is 2 x 2; the 6 th layer and the 7 th layer are full-connection layers, and are respectively 120 nerve nodes and 84 nerve nodes; the 8 th layer is an output layer, and the output node is N.
Further, in step S3, the convolution kernels of the second layer and the fourth layer of the deformable convolution network structure discard part of the convolution kernel parameters according to the following rule:
(1) converting each picture of the training set into a gray level image, converting the gray level image into a binary image, compressing the binary image of 90 x 90 to a 9*9 gray level image, and obtaining the binary image;
(2) adding all 9*9 binary pictures in the step (1), wherein the white point value of the binary pictures is 1, and the black point value of the binary pictures is 0; specifically, pixel values of positions corresponding to each pixel point of the picture are added and divided by the number of the training set pictures, and each pixel point is multiplied by 255 to obtain an average Gray image Gray of 9*9;
(3) the average Gray image Gray average Mean of the Gray images Gray is calculated, and then binarization operation is carried out on the average Gray images Gray by taking Mean as a threshold value to obtain a binary image C of 9*9;
(4) the 9*9 convolution kernel parameter discarding portions of the second layer and the fourth layer correspond to pixel positions with a pixel value of 0 in the 9*9 binary image C.
Further, the binarization operation adopts an Ojin binarization method.
Further, the image scaling adopts a reserve () function of opencv to select a nearest neighbor interpolation mode.
On the other hand, the invention also provides a deep learning image classification system based on the deformable convolution network, which comprises the following steps:
the image scaling module is used for scaling all the images to 90 x 90 and dividing the images into a training set and a testing set;
the network structure setting module is used for setting a deformable convolution network structure;
the parameter discarding module is used for discarding part of convolution kernel parameters of the second layer and the fourth layer of the deformable convolution network structure according to a certain rule:
the model training module is used for training the training set data according to the network model;
and the classification and identification module is used for classifying and identifying the test set data by the trained model.
Further, the deformable convolutional network structure is specifically: layer 1 is an input layer, and 90 x 90 pictures are input; the layer 2 is a convolution layer, the number of convolution kernels is 16, and the size of the convolution kernels is 9*9; layer 3 is the maximum pooling layer, with a core size of 2 x 2; the 4 th layer is a convolution layer, the number of convolution kernels is 16, and the size of the convolution kernels is 9*9; layer 5 is the maximum pooling layer, the core size is 2 x 2; the 6 th layer and the 7 th layer are full-connection layers, and are respectively 120 nerve nodes and 84 nerve nodes; the 8 th layer is an output layer, and the output node is N.
Further, the convolution kernels of the second and fourth layers of the deformable convolution network structure discard a portion of the convolution kernel parameters according to the following rules:
(1) converting each picture of the training set into a gray level image, converting the gray level image into a binary image, compressing the binary image of 90 x 90 to a 9*9 gray level image, and obtaining the binary image;
(2) adding all 9*9 binary pictures in the step (1), wherein the white point value of the binary pictures is 1, and the black point value of the binary pictures is 0; specifically, pixel values of positions corresponding to each pixel point of the picture are added and divided by the number of the training set pictures, and each pixel point is multiplied by 255 to obtain an average Gray image Gray of 9*9;
(3) the average Gray image Gray average Mean of the Gray images Gray is calculated, and then binarization operation is carried out on the average Gray images Gray by taking Mean as a threshold value to obtain a binary image C of 9*9;
(4) the 9*9 convolution kernel parameter discarding portions of the second layer and the fourth layer correspond to pixel positions with a pixel value of 0 in the 9*9 binary image C.
Further, the binarization operation adopts an Ojin binarization method.
Further, the image scaling adopts a reserve () function of opencv to select a nearest neighbor interpolation mode.
Compared with the prior art, the invention has the advantages and positive effects that at least the following steps are included:
(1) According to the method, the convolution kernel is automatically designed according to the characteristic information of the classified images, and partial parameters are discarded, so that the network operation calculation is reduced and the network operation efficiency is improved while the network key characteristic information extraction capability is improved;
(2) The method is suitable for other image classification tasks and has certain universality.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a deformable convolutional network-based deep learning image classification method of the present invention;
FIG. 2 is a schematic diagram of a network structure used in the deep learning image classification method based on a deformable convolutional network of the present invention;
fig. 3 is a schematic structural diagram of a deep learning image classification system based on a deformable convolution network according to the present invention.
Detailed Description
In order to make the above objects, features and advantages of the present invention more comprehensible, the following detailed description of the technical solution of the present invention refers to the accompanying drawings and specific embodiments. It should be noted that the described embodiments are only some embodiments of the present invention, and not all embodiments, and that all other embodiments obtained by persons skilled in the art without making creative efforts based on the embodiments in the present invention are within the protection scope of the present invention.
Example 1
To illustrate the present invention, the present embodiment takes CIFAR-10 dataset as an example of a 10-class problem.
It should be noted that, the specific numbers used in the present embodiment are merely a set of possible or preferred combinations used in the present embodiment, and should not be construed as limiting the scope of the invention; it should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the invention, which are all within the scope of the invention. Accordingly, the scope of the invention should be assessed as that of the appended claims.
As shown in fig. 1, the present embodiment provides a deep learning image classification method based on a deformable convolution network, which includes the following steps:
step S1: the CIFAR-10 dataset has a total of 10 pictures, including (airplane, car, bird, cat, deer, dog, frog, horse, boat, and truck), scaling all images to 90 x 90 size, and separating into training and test sets;
step S2: setting a network structure: layer 1 is an input layer, and 90 x 90 pictures are input; the layer 2 is a convolution layer, the number of convolution kernels is 16, and the size of the convolution kernels is 9*9; layer 3 is the maximum pooling layer, with a core size of 2 x 2; the 4 th layer is a convolution layer, the number of convolution kernels is 16, and the size of the convolution kernels is 9*9; layer 5 is the maximum pooling layer, the core size is 2 x 2; the 6 th layer and the 7 th layer are full-connection layers, and are respectively 120 nerve nodes and 84 nerve nodes; layer 8 is the output layer and the output node is N, as shown in fig. 2.
Step S3: in step S2, the convolution kernels of the second layer and the fourth layer discard part of the convolution kernel parameters according to the following rule:
(1) converting each picture of the training set into a gray level image, converting the gray level image into a binary image, compressing the binary image of 90 x 90 to a 9*9 gray level image, and obtaining the binary image;
(2) adding all 9*9 binary pictures (white point value is 1, black point value is 0) in the step (1), specifically adding pixel values at positions corresponding to each pixel point of the pictures, dividing the added pixel values by the number of training set pictures, and multiplying each pixel point by 255 to obtain an average Gray level graph Gray of 9*9;
(3) the average Gray image Gray average Mean of the Gray images Gray is calculated, and then binarization operation is carried out on the average Gray images Gray by taking Mean as a threshold value to obtain a binary image C of 9*9;
(4) the 9*9 convolution kernel parameter discarding portions of the second layer and the fourth layer correspond to pixel positions with a pixel value of 0 in the 9*9 binary image C.
Step S4: training the training set data according to the network model;
step S5: and carrying out classification recognition on the test set data by the trained model to obtain a classification recognition result.
The method is based on the comparison test of the CIFAR-10 data set and three mainstream deep learning algorithms of AlexNet, VGG16 and GoogleNet.
Table 1 comparison of test results of four methods
| Method | Accuracy rate of |
| The method of the invention | 98.83％ |
| AlexNet | 92.34％ |
| VGG16 | 96.71％ |
| GoogleNet | 97.45％ |
As can be seen from the table 1, the accuracy of the method is higher than that of the existing mainstream deep learning networks AlexNet, VGG16 and GoogleNet.
Example 2
To illustrate the present invention, the present embodiment takes CIFAR-10 dataset as an example of a 10-class problem.
As shown in fig. 3, the embodiment provides a deep learning image classification system based on a deformable convolution network, which comprises an image scaling module, a network structure setting module, a parameter discarding module, a model training module and a classification recognition module;
an image scaling module: the CIFAR-10 dataset has a total of 10 pictures, including (airplane, car, bird, cat, deer, dog, frog, horse, boat, and truck), scaling all images to 90 x 90 size, and separating into training and test sets;
a network structure setting module: setting a network structure, wherein the 1 st layer is an input layer, and inputting 90-90 pictures; the layer 2 is a convolution layer, the number of convolution kernels is 16, and the size of the convolution kernels is 9*9; layer 3 is the maximum pooling layer, with a core size of 2 x 2; the 4 th layer is a convolution layer, the number of convolution kernels is 16, and the size of the convolution kernels is 9*9; layer 5 is the maximum pooling layer, the core size is 2 x 2; the 6 th layer and the 7 th layer are full-connection layers, and are respectively 120 nerve nodes and 84 nerve nodes; layer 8 is the output layer and the output node is N, as shown in fig. 2.
And the parameter discarding module is used for: discarding part of convolution kernel parameters of the second layer and the fourth layer of the network structure according to the following rule:
(1) converting each picture of the training set into a gray level image, converting the gray level image into a binary image, compressing the binary image of 90 x 90 to a 9*9 gray level image, and obtaining the binary image;
(2) adding all 9*9 binary pictures (white point value is 1, black point value is 0) in the step (1), specifically adding pixel values at positions corresponding to each pixel point of the pictures, dividing the added pixel values by the number of training set pictures, and multiplying each pixel point by 255 to obtain an average Gray level graph Gray of 9*9;
(3) the average Gray image Gray average Mean of the Gray images Gray is calculated, and then binarization operation is carried out on the average Gray images Gray by taking Mean as a threshold value to obtain a binary image C of 9*9;
(4) the 9*9 convolution kernel parameter discarding portions of the second layer and the fourth layer correspond to pixel positions with a pixel value of 0 in the 9*9 binary image C.
Model training module: training the training set data according to the network model;
the classification and identification module: and carrying out classification recognition on the test set data by the trained model to obtain a classification recognition result.
The invention has the advantages and positive effects that at least comprises:
(1) According to the method, the convolution kernel is automatically designed according to the characteristic information of the classified images, and partial parameters are discarded, so that the network operation calculation is reduced and the network operation efficiency is improved while the network key characteristic information extraction capability is improved;
(2) The method is suitable for other image classification tasks and has certain universality.
The foregoing examples illustrate only a few embodiments of the invention and are described in detail herein without thereby limiting the scope of the invention. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the invention, which are all within the scope of the invention. Accordingly, the scope of the invention should be assessed as that of the appended claims.
Claims (8)
1. The deep learning image classification method based on the deformable convolution network is characterized by comprising the following steps of:
step S1: scaling all images to 90 x 90 size and dividing them into training set and test set;
step S2: setting a deformable convolution network structure;
step S3: discarding part of convolution kernel parameters of the second layer and the fourth layer of the deformable convolution network structure according to a certain rule, wherein the method specifically comprises the following steps:
(1) converting each picture of the training set into a gray level image, converting the gray level image into a binary image, compressing the binary image of 90 x 90 to a 9*9 gray level image, and obtaining the binary image;
(2) adding all 9*9 binary pictures in the step (1), wherein the white point value of the binary pictures is 1, and the black point value of the binary pictures is 0; specifically, pixel values of positions corresponding to each pixel point of the picture are added and divided by the number of the training set pictures, and each pixel point is multiplied by 255 to obtain an average Gray image Gray of 9*9;
(3) the average Gray image Gray average Mean of the Gray images Gray is calculated, and then binarization operation is carried out on the average Gray images Gray by taking Mean as a threshold value to obtain a binary image C of 9*9;
(4) the 9*9 convolution kernel parameter discarding parts of the second layer and the fourth layer correspond to pixel positions with the pixel value of 0 in the 9*9 binary image C;
step S4: training the training set data according to the network model;
step S5: and classifying and identifying the test set data by the trained model.
2. The deep learning image classification method based on the deformable convolution network according to claim 1, wherein in step S2, the deformable convolution network structure specifically comprises: layer 1 is an input layer, and 90 x 90 pictures are input; the layer 2 is a convolution layer, the number of convolution kernels is 16, and the size of the convolution kernels is 9*9; layer 3 is the maximum pooling layer, with a core size of 2 x 2; the 4 th layer is a convolution layer, the number of convolution kernels is 16, and the size of the convolution kernels is 9*9; layer 5 is the maximum pooling layer, the core size is 2 x 2; the 6 th layer and the 7 th layer are full-connection layers, and are respectively 120 nerve nodes and 84 nerve nodes; the 8 th layer is an output layer, and the output node is N.
3. The method for classifying deep learning images based on a deformable convolutional network according to claim 1, wherein the binarization operation adopts an oxford binarization method.
4. The method for classifying deep learning images based on a deformable convolutional network according to claim 1, wherein the image scaling uses a restore () function of opencv to select nearest neighbor interpolation mode.
5. A deep learning image classification system based on a deformable convolutional network, comprising:
the image scaling module is used for scaling all the images to 90 x 90 and dividing the images into a training set and a testing set;
the network structure setting module is used for setting a deformable convolution network structure;
the parameter discarding module is configured to discard a part of convolution kernel parameters of the second layer and the fourth layer of the deformable convolution network structure according to a certain rule, and specifically includes:
(1) converting each picture of the training set into a gray level image, converting the gray level image into a binary image, compressing the binary image of 90 x 90 to a 9*9 gray level image, and obtaining the binary image;
(2) adding all 9*9 binary pictures in the step (1), wherein the white point value of the binary pictures is 1, and the black point value of the binary pictures is 0; specifically, pixel values of positions corresponding to each pixel point of the picture are added and divided by the number of the training set pictures, and each pixel point is multiplied by 255 to obtain an average Gray image Gray of 9*9;
(3) the average Gray image Gray average Mean of the Gray images Gray is calculated, and then binarization operation is carried out on the average Gray images Gray by taking Mean as a threshold value to obtain a binary image C of 9*9;
(4) the 9*9 convolution kernel parameter discarding parts of the second layer and the fourth layer correspond to pixel positions with the pixel value of 0 in the 9*9 binary image C;
the model training module is used for training the training set data according to the network model;
and the classification and identification module is used for classifying and identifying the test set data by the trained model.
6. The deep learning image classification system based on a deformable convolutional network of claim 5, wherein the deformable convolutional network structure is specifically: layer 1 is an input layer, and 90 x 90 pictures are input; the layer 2 is a convolution layer, the number of convolution kernels is 16, and the size of the convolution kernels is 9*9; layer 3 is the maximum pooling layer, with a core size of 2 x 2; the 4 th layer is a convolution layer, the number of convolution kernels is 16, and the size of the convolution kernels is 9*9; layer 5 is the maximum pooling layer, the core size is 2 x 2; the 6 th layer and the 7 th layer are full-connection layers, and are respectively 120 nerve nodes and 84 nerve nodes; the 8 th layer is an output layer, and the output node is N.
7. The deformable convolutional network-based deep learning image classification system of claim 6, wherein the binarization operation employs an oxford binarization method.
8. The deformable convolutional network-based deep learning image classification system of claim 7, wherein the image scaling employs opencv's resize () function to select nearest neighbor interpolation mode.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110303551.5A CN113139577B (en) | 2021-03-22 | 2021-03-22 | Deep learning image classification method and system based on deformable convolution network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110303551.5A CN113139577B (en) | 2021-03-22 | 2021-03-22 | Deep learning image classification method and system based on deformable convolution network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113139577A CN113139577A (en) | 2021-07-20 |
| CN113139577B true CN113139577B (en) | 2024-02-23 |
Family
ID=76810271
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110303551.5A Active CN113139577B (en) | 2021-03-22 | 2021-03-22 | Deep learning image classification method and system based on deformable convolution network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113139577B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114022714B (en) * | 2021-11-11 | 2024-04-16 | 哈尔滨工程大学 | Harris-based data enhanced image classification method and system |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108510467A (en) * | 2018-03-28 | 2018-09-07 | 西安电子科技大学 | SAR image target recognition method based on variable depth shape convolutional neural networks |
| CN109086799A (en) * | 2018-07-04 | 2018-12-25 | 江苏大学 | A kind of crop leaf disease recognition method based on improvement convolutional neural networks model AlexNet |
| CN109522949A (en) * | 2018-11-07 | 2019-03-26 | 北京交通大学 | Model of Target Recognition method for building up and device |
| CN109871799A (en) * | 2019-02-02 | 2019-06-11 | 浙江万里学院 | A kind of driver based on deep learning plays the detection method of mobile phone behavior |
| CN110222817A (en) * | 2019-05-10 | 2019-09-10 | 上海交通大学 | Convolutional neural networks compression method, system and medium based on learning automaton |
| CN110427922A (en) * | 2019-09-03 | 2019-11-08 | 陈�峰 | One kind is based on machine vision and convolutional neural networks pest and disease damage identifying system and method |
| CN110598731A (en) * | 2019-07-31 | 2019-12-20 | 浙江大学 | Efficient image classification method based on structured pruning |
| CN111027626A (en) * | 2019-12-11 | 2020-04-17 | 西安电子科技大学 | Flow field identification method based on deformable convolution network |
| CN111368684A (en) * | 2020-02-27 | 2020-07-03 | 北华航天工业学院 | Winter wheat automatic interpretation method based on deformable full-convolution neural network |
| CN112465810A (en) * | 2020-12-15 | 2021-03-09 | 华南农业大学 | Method for detecting and classifying defects of textiles |
-
2021
- 2021-03-22 CN CN202110303551.5A patent/CN113139577B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108510467A (en) * | 2018-03-28 | 2018-09-07 | 西安电子科技大学 | SAR image target recognition method based on variable depth shape convolutional neural networks |
| CN109086799A (en) * | 2018-07-04 | 2018-12-25 | 江苏大学 | A kind of crop leaf disease recognition method based on improvement convolutional neural networks model AlexNet |
| CN109522949A (en) * | 2018-11-07 | 2019-03-26 | 北京交通大学 | Model of Target Recognition method for building up and device |
| CN109871799A (en) * | 2019-02-02 | 2019-06-11 | 浙江万里学院 | A kind of driver based on deep learning plays the detection method of mobile phone behavior |
| CN110222817A (en) * | 2019-05-10 | 2019-09-10 | 上海交通大学 | Convolutional neural networks compression method, system and medium based on learning automaton |
| CN110598731A (en) * | 2019-07-31 | 2019-12-20 | 浙江大学 | Efficient image classification method based on structured pruning |
| CN110427922A (en) * | 2019-09-03 | 2019-11-08 | 陈�峰 | One kind is based on machine vision and convolutional neural networks pest and disease damage identifying system and method |
| CN111027626A (en) * | 2019-12-11 | 2020-04-17 | 西安电子科技大学 | Flow field identification method based on deformable convolution network |
| CN111368684A (en) * | 2020-02-27 | 2020-07-03 | 北华航天工业学院 | Winter wheat automatic interpretation method based on deformable full-convolution neural network |
| CN112465810A (en) * | 2020-12-15 | 2021-03-09 | 华南农业大学 | Method for detecting and classifying defects of textiles |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113139577A (en) | 2021-07-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110992275B (en) | Refined single image rain removing method based on generation of countermeasure network | |
| CN110555372A (en) | Data entry method, device, equipment and storage medium | |
| CN107292458B (en) | Prediction method and prediction device applied to neural network chip | |
| CN109410146A (en) | A kind of image deblurring algorithm based on Bi-Skip-Net | |
| JP6341650B2 (en) | Image processing apparatus, image processing method, and program | |
| JP2008542911A (en) | Image comparison by metric embedding | |
| CN111274987B (en) | Facial expression recognition method and facial expression recognition device | |
| JP2014041476A (en) | Image processing apparatus, image processing method, and program | |
| CN113781510B (en) | Edge detection method and device and electronic equipment | |
| CN109740553B (en) | Image semantic segmentation data screening method and system based on recognition | |
| JP6107531B2 (en) | Feature extraction program and information processing apparatus | |
| US11605210B2 (en) | Method for optical character recognition in document subject to shadows, and device employing method | |
| CN112560901A (en) | Method for defending and confronting sample based on combination of image preprocessing and confronting training | |
| CN110751191A (en) | Image classification method and system | |
| CN113139577B (en) | Deep learning image classification method and system based on deformable convolution network | |
| CN105354228B (en) | Similar diagram searching method and device | |
| CN113971644A (en) | Image identification method and device based on data enhancement strategy selection | |
| CN116363535A (en) | Ship detection method in unmanned aerial vehicle aerial image based on convolutional neural network | |
| CN109003247B (en) | Method for removing color image mixed noise | |
| CN110728352A (en) | Large-scale image classification method based on deep convolutional neural network | |
| CN110288604B (en) | Image segmentation method and device based on K-means | |
| CN110349119B (en) | Pavement disease detection method and device based on edge detection neural network | |
| CN113988148A (en) | Data clustering method, system, computer equipment and storage medium | |
| CN116091844A (en) | Image data processing method and system based on edge calculation | |
| CN116245765A (en) | Image denoising method and system based on enhanced depth expansion convolutional neural network |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |