CN114444653A - Method and system for evaluating influence of data augmentation on deep learning model performance - Google Patents
Method and system for evaluating influence of data augmentation on deep learning model performance Download PDFInfo
- Publication number
- CN114444653A CN114444653A CN202111576340.5A CN202111576340A CN114444653A CN 114444653 A CN114444653 A CN 114444653A CN 202111576340 A CN202111576340 A CN 202111576340A CN 114444653 A CN114444653 A CN 114444653A
- Authority
- CN
- China
- Prior art keywords
- data set
- augmentation
- images
- deep learning
- learning model
- 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
- 238000013136 deep learning model Methods 0.000 title claims abstract description 47
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000013434 data augmentation Methods 0.000 title claims abstract description 26
- 230000003416 augmentation Effects 0.000 claims abstract description 44
- 238000012549 training Methods 0.000 claims abstract description 23
- 238000012360 testing method Methods 0.000 claims abstract description 15
- 230000003321 amplification Effects 0.000 claims abstract description 11
- 238000003199 nucleic acid amplification method Methods 0.000 claims abstract description 11
- 238000011156 evaluation Methods 0.000 claims abstract description 7
- 230000003190 augmentative effect Effects 0.000 claims description 47
- 238000013508 migration Methods 0.000 claims description 31
- 230000005012 migration Effects 0.000 claims description 31
- 238000004590 computer program Methods 0.000 claims description 12
- 238000013135 deep learning Methods 0.000 abstract description 5
- 238000004458 analytical method Methods 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010835 comparative analysis Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000010008 shearing Methods 0.000 description 1
- 230000005477 standard model Effects 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- 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
- 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
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/04—Context-preserving transformations, e.g. by using an importance map
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)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a method and a system for evaluating the influence of data augmentation on the performance of a deep learning model, wherein the method comprises the following steps: acquiring an original data set of a deep learning model; based on a generative countermeasure network, amplifying the original data set through different amplification modes to obtain a plurality of amplification data sets; and training the deep learning model by using each augmentation data set respectively, and testing and evaluating each trained deep learning model. The deep learning network is trained and tested through the initial data set to obtain the standard evaluation index, so that the subsequent comparison and analysis of the model training test index based on the data augmentation of the confrontation network are facilitated.
Description
Technical Field
The invention belongs to the technical field of deep learning and visual recognition, and particularly relates to a method and a system for evaluating the influence of data augmentation on the performance of a deep learning model.
Background
Data augmentation is one of the common skills in deep learning, and is mainly used for increasing a training data set, so that the data set is diversified as much as possible, and a trained model has stronger generalization capability. The current data augmentation mainly comprises: the method comprises the following steps of horizontal/vertical turning, rotation, zooming, clipping, shearing, translation, contrast, color dithering, noise and the like, wherein the change premises are that the label of an image is not changed and only limited in the field of the image, and the traditional data augmentation algorithms have small influence on the performance of a deep learning model and do not fundamentally solve the problem of insufficient data.
Some recently emerging generative confrontation network models attract a lot of attention due to their excellent performance, and compared with the conventional data enhancement technology, the network synthesis-based method has a more complicated process, but the generated samples are more diverse.
Disclosure of Invention
In order to solve the problem of influence of augmented data on a model and verify and determine an optimal data augmentation mode, the invention provides a method for evaluating the influence of data augmentation on the performance of a deep learning model in a first aspect, which comprises the following steps: acquiring an original data set of a deep learning model; based on a generative countermeasure network, amplifying the original data set through different amplification modes to obtain a plurality of amplification data sets; and training the deep learning model by using each augmentation data set respectively, and testing and evaluating each trained deep learning model.
In some embodiments of the present invention, the augmenting the original data set by different augmenting manners based on the generative countermeasure network to obtain a plurality of augmented data sets includes: performing style migration on one or more images in the original data set based on the generative countermeasure network to obtain a first augmented data set; performing target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a second augmented data set; and performing style migration or target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a third augmented data set.
Further, the performing style migration on one or more images in the original data set based on the generative countermeasure network to obtain a first augmented data set includes: taking one or more images in the original data set as target images; and migrating the style of the image in the non-original data set to the target image by using a generative countermeasure network to obtain one or more augmented images.
Further, the performing target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a second augmented data set includes: taking one or more images in the original data set as target images; and replacing one or more targets in the images in the non-original data set with corresponding targets in the target images by using the generative countermeasure network to obtain one or more augmented images.
Further, the performing style migration or target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a third augmented data set includes: and fusing the images in the first and second augmented data sets to obtain a third augmented data set.
In the above embodiment, the generative countermeasure network is MSGAN.
In a second aspect of the present invention, there is provided a system for evaluating the influence of data augmentation on the performance of a deep learning model, comprising: the acquisition module is used for acquiring an original data set of the deep learning model; the augmentation module is used for augmenting the original data set based on the generative countermeasure network in different augmentation modes to obtain a plurality of augmentation data sets; and the evaluation module is used for training the deep learning model by utilizing each augmentation data set respectively and testing and evaluating each trained deep learning model.
Further, the augmentation module includes: the first augmentation unit is used for carrying out style migration on one or more images in the original data set based on the generative countermeasure network to obtain a first augmentation data set; the second augmentation unit is used for carrying out target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a second augmentation data set; and the third augmentation unit is used for carrying out style migration or target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a third augmentation data set.
In a third aspect of the present invention, there is provided an electronic device comprising: one or more processors; a storage device for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to implement the method for evaluating the impact of data augmentation on the performance of a deep learning model provided by the invention in the first aspect.
In a fourth aspect of the present invention, a computer-readable medium is provided, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the method for evaluating the influence of data augmentation on the deep learning model performance provided by the first aspect of the present invention.
The invention has the beneficial effects that:
1. the deep learning network is trained and tested through the initial data set to obtain a standard evaluation index, so that the subsequent model training test index based on the data augmentation of the confrontation network is convenient to compare and analyze, and the model is trained by three data augmentation modes, namely picture style migration, target replacement and target replacement + picture style migration, and is most effective in comparing which data augmentation mode is the most effective;
2. according to the method, through the comparative analysis among different models, the style migration is found to be based on the whole picture to modify the picture color, and the target replacement is based on the local target modification, so that the integration of two augmentation algorithms is most effective for improving the indexes of the models.
Drawings
FIG. 1 is a schematic flow chart of a method for evaluating the impact of data augmentation on deep learning model performance in some embodiments of the present invention;
FIG. 2 is a flowchart illustrating a method for evaluating the impact of data augmentation on deep learning model performance in some embodiments of the present invention;
FIG. 3 is a schematic diagram illustrating a training data set augmentation method based on a generative confrontation network in some embodiments of the present invention;
fig. 4 is a schematic diagram illustrating a contextual comparison of an image in a source data set with a target replaced in accordance with some embodiments of the present invention;
FIG. 5 is a schematic diagram illustrating a contextual migration of images in a source data set according to some embodiments of the invention;
FIG. 6 is a diagram illustrating results of calls under different algorithms with various arrows in some embodiments of the present invention;
FIG. 7 is an arrow model index of different augmentation methods of style migration in some embodiments of the present invention
FIG. 8 is a block diagram of a system for evaluating the impact of data augmentation on deep learning model performance in some embodiments of the present invention;
fig. 9 is a schematic structural diagram of an electronic device in some embodiments of the invention.
Detailed Description
The principles and features of this invention are described below in conjunction with the following drawings, which are set forth by way of illustration only and are not intended to limit the scope of the invention.
Referring to fig. 1 and fig. 2, in a first aspect of the present invention, a method for evaluating an influence of data augmentation on a deep learning model performance is provided, including: s100, acquiring an original data set of a deep learning model; s200, amplifying the original data set based on a generating type countermeasure network through different amplification modes to obtain a plurality of amplification data sets; s300, training the deep learning model by using each augmentation data set respectively, and testing and evaluating each trained deep learning model.
In order to accelerate the convergence rate of the deep learning model, in some embodiments of the present invention, the method further includes: initializing a deep learning model before obtaining an original data set of the deep learning model, wherein the method specifically comprises the following steps:
step A: in order to verify that the effectiveness of the deep learning algorithm can be improved after the training data set is augmented by using the countermeasure network, the method takes four arrows of a straight arrow, a turning arrow, a left-turn or turning arrow and a straight or turning arrow in an urban road as an example, and under a natural scene state, the number ratio of the four arrows is 14: 6: 3: 2, it can be seen from this ratio that the distribution ratios of different arrows are greatly different, and when the model is trained by using such data, the generalization capability of the model is not high.
And B, step B: training an initialized model using the raw data set of S11 can be used as a reference model for subsequent training using the augmented data set as a comparison experiment. The deep learning model is yolov5s, the training method is to load the labeled training data set as input into the yolov5s network model, and obtain an optimal algorithm model 1 after multiple iterations, and the strategy for selecting the optimal model is calculated based on the map index evaluated by the target detection index.
And C: the test indexes of the model are evaluated based on a test data set (different from a training data set), and according to an index evaluation algorithm of target detection, precision, recall and map indexes of four arrows can be obtained respectively, and meanwhile precision, recall and map values of all test targets also need to be given. A schematic diagram of a specific model training test is shown in fig. 2. Optionally, the deep learning model selected in the present invention further includes: an object detection model, a semantic segmentation model, or an instance segmentation model, etc.
Referring to fig. 3, in some embodiments of the present invention, the augmenting the original data set by different augmenting manners based on the generative countermeasure network to obtain a plurality of augmented data sets includes: s201, carrying out style migration on one or more images in the original data set based on a generating type countermeasure network to obtain a first augmented data set; s202, performing target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a second augmented data set; and S203, performing style migration or target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a third augmented data set.
Further, in step S201, the performing style migration on one or more images in the original data set based on the generative countermeasure network to obtain a first augmented data set includes: taking one or more images in the original data set as target images; and migrating the style of the image in the non-original data set to the target image by using a generative countermeasure network to obtain one or more augmented images.
Specifically, the style migration network is used for converting road pictures in natural scenes under different scenes, the scene roads of the invention include cloudy roads, rainy roads, roads under strong sunlight, roads at night and the like, and the specific style migration effect graph is shown in fig. 4. And taking one picture in the training data set as a content picture, randomly selecting one picture in different road scene images as a style picture, finally sending the content picture and the style picture into an image style migration network MSGAN to obtain a content picture in other styles as an output picture, and converting other training data sets in the same way, so that the number of the training data sets can be doubled, and a new evaluation index of the model 2 can be obtained by using the augmented data set according to the model training and testing method in the steps B-C. Fig. 4 shows the migration of a road image with insufficient light to a road image with sufficient light on the cloudy day.
Further, in step S202, the performing target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a second augmented data set includes: taking one or more images in the original data set as target images; and replacing one or more targets in the images in the non-original data set with corresponding targets in the target images by using the generative countermeasure network to obtain one or more augmented images. Schematically, fig. 5 shows images before and after the replacement, wherein the image of the left half is derived from the original data set, and after the replacement, the straight arrow in the road becomes a arrow with a bifurcation or a turn arrow as can be seen from the image in the rectangular framed area.
Further, in step S203, the performing style migration or target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a third augmented data set includes: and fusing the images in the first and second augmented data sets to obtain a third augmented data set.
In the above embodiment, the Generative countermeasure network is msgan (model generating adaptive networks).
In the above embodiment, the deep learning model is trained using each augmented data set, and each trained deep learning model is tested and evaluated. Specifically, the method comprises the following steps:
s301, through the three different modes of the arrow training data set, from the test indexes, the precision values of the different algorithm amplification indexes are not changed greatly, so that the experiment analyzes the call indexes of the different amplification algorithms to obtain the result shown in FIG. 6.
S302: because the image migration (image migration) is based on all images (images) to carry out style transformation, the arrow recall indexes of all types are improved to a certain extent compared with the standard model indexes; since the target replacement is to replace a part of the straight arrow with the straight or u-turn arrow and the left-turn or u-turn arrow, it can be seen from fig. 5 that the call value of the straight arrow is decreased, and the call value of the straight or u-turn arrow and the left-turn or u-turn arrow is increased by about 15% compared with the reference model, so that it can be illustrated that a scheme of replacing a small number of targets with a large number of targets is feasible.
S303: it can be seen from fig. 7 that the model performance of the three different augmentation algorithms is: the three augmentation modes are relatively slow in improvement of precision indexes and relatively large in improvement of recall and map indexes, wherein in the augmentation mode of replacement and migration, the recall index is improved by 12% compared with a reference model, and the model is very effective in a training process. Each augmentation algorithm is effective, but the effectiveness of the different augmentation algorithms is different, and since the style migration is based on modifying the picture color of the whole picture and the target replacement is based on modifying the local target, the combination of the two augmentation algorithms is most effective for improving the index of the model.
Example 2
Referring to fig. 8, in a second aspect of the present invention, there is provided a system 1 for evaluating influence of data augmentation on deep learning model performance, comprising: the acquisition module 11 is used for acquiring an original data set of the deep learning model; an augmentation module 12, configured to augment the original data set in different augmentation modes based on a generative countermeasure network to obtain multiple augmentation data sets; and the evaluation module 13 is configured to train the deep learning model by using each augmented data set, and test and evaluate each trained deep learning model.
Further, the amplification module 12 includes: the first augmentation unit is used for carrying out style migration on one or more images in the original data set based on the generative countermeasure network to obtain a first augmentation data set; the second augmentation unit is used for carrying out target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a second augmentation data set; and the third augmentation unit is used for carrying out style migration or target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a third augmentation data set.
Example 3
Referring to fig. 9, in a third aspect of the present invention, there is provided an electronic apparatus comprising: one or more processors; storage means for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to carry out the method of the invention in the first aspect.
The electronic device 500 may include a processing means (e.g., central processing unit, graphics processor, etc.) 501 that may perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM)502 or a program loaded from a storage means 508 into a Random Access Memory (RAM) 503. In the RAM 503, various programs and data necessary for the operation of the electronic apparatus 500 are also stored. The processing device 501, the ROM502, and the RAM 503 are connected to each other through a bus 504. An input/output (I/O) interface 505 is also connected to bus 504.
The following devices may be connected to the I/O interface 505 in general: input devices 506 including, for example, a touch screen, touch pad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; output devices 507 including, for example, a Liquid Crystal Display (LCD), speakers, vibrators, and the like; a storage device 508 including, for example, a hard disk; and a communication device 509. The communication means 509 may allow the electronic device 500 to communicate with other devices wirelessly or by wire to exchange data. While fig. 9 illustrates an electronic device 500 having various means, it is to be understood that not all illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided. Each block shown in fig. 9 may represent one device or may represent multiple devices as desired.
In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network via the communication means 509, or installed from the storage means 508, or installed from the ROM 502. The computer program, when executed by the processing device 501, performs the above-described functions defined in the methods of embodiments of the present disclosure. It should be noted that the computer readable medium described in the embodiments of the present disclosure may be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In embodiments of the disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In embodiments of the present disclosure, however, a computer readable signal medium may comprise a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, optical cables, RF (radio frequency), etc., or any suitable combination of the foregoing.
The computer readable medium may be embodied in the electronic device; or may exist separately without being assembled into the electronic device. The computer readable medium carries one or more computer programs which, when executed by the electronic device, cause the electronic device to:
computer program code for carrying out operations for embodiments of the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + +, Python, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (10)
1. A method for evaluating the influence of data augmentation on the performance of a deep learning model is characterized by comprising the following steps:
acquiring an original data set of a deep learning model;
based on a generative countermeasure network, amplifying the original data set through different amplification modes to obtain a plurality of amplification data sets;
and training the deep learning model by using each augmentation data set respectively, and testing and evaluating each trained deep learning model.
2. The method of claim 1, wherein the augmenting the raw data set with different augmenting methods based on the generative countermeasure network to augment the raw data set to obtain a plurality of augmented data sets comprises:
performing style migration on one or more images in the original data set based on the generative countermeasure network to obtain a first augmented data set;
performing target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a second augmented data set;
and performing style migration or target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a third augmented data set.
3. The method of claim 2, wherein the performing a style migration on one or more images in the original data set based on the generative confrontation network to obtain a first augmented data set comprises:
taking one or more images in the original data set as target images;
and migrating the style of the image in the non-original data set to the target image by using a generative countermeasure network to obtain one or more augmented images.
4. The method of claim 2, wherein the performing target replacement on one or more images in the original data set based on the generative confrontation network to obtain a second augmented data set comprises:
taking one or more images in the original data set as target images;
and replacing one or more targets in the images in the non-original data set with corresponding targets in the target images by using the generative countermeasure network to obtain one or more augmented images.
5. The method of claim 2, wherein the performing style migration or target replacement on one or more images in the original data set based on the generative confrontation network to obtain a third augmented data set comprises:
and fusing the images in the first and second augmented data sets to obtain a third augmented data set.
6. The method for evaluating the influence of data augmentation on the performance of the deep learning model according to any one of claims 1 to 5, wherein the generative confrontation network is MSGAN.
7. A system for assessing the impact of data augmentation on the performance of a deep learning model, comprising:
the acquisition module is used for acquiring an original data set of the deep learning model;
the augmentation module is used for augmenting the original data set based on the generative countermeasure network in different augmentation modes to obtain a plurality of augmentation data sets;
and the evaluation module is used for training the deep learning model by utilizing each augmentation data set respectively and testing and evaluating each trained deep learning model.
8. The system of claim 7, wherein the augmentation module comprises:
the first augmentation unit is used for carrying out style migration on one or more images in the original data set based on the generative countermeasure network to obtain a first augmentation data set;
the second augmentation unit is used for carrying out target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a second augmentation data set;
and the third augmentation unit is used for carrying out style migration or target replacement on one or more images in the original data set based on the generative countermeasure network to obtain a third augmentation data set.
9. An electronic device, comprising: one or more processors; a storage device for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to implement the method for estimating impact of data augmentation on deep learning model performance as claimed in any one of claims 1 to 6.
10. A computer-readable medium, on which a computer program is stored, wherein the computer program, when being executed by a processor, carries out a method for estimating an impact of data augmentation on deep learning model performance as defined in any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111576340.5A CN114444653A (en) | 2021-12-21 | 2021-12-21 | Method and system for evaluating influence of data augmentation on deep learning model performance |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111576340.5A CN114444653A (en) | 2021-12-21 | 2021-12-21 | Method and system for evaluating influence of data augmentation on deep learning model performance |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114444653A true CN114444653A (en) | 2022-05-06 |
Family
ID=81363898
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111576340.5A Pending CN114444653A (en) | 2021-12-21 | 2021-12-21 | Method and system for evaluating influence of data augmentation on deep learning model performance |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114444653A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115273072A (en) * | 2022-06-13 | 2022-11-01 | 南京林业大学 | Apple leaf disease detection method based on improved Yolov5s model |
| CN115345321A (en) * | 2022-10-19 | 2022-11-15 | 小米汽车科技有限公司 | Data augmentation method, data augmentation device, electronic device, and storage medium |
-
2021
- 2021-12-21 CN CN202111576340.5A patent/CN114444653A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115273072A (en) * | 2022-06-13 | 2022-11-01 | 南京林业大学 | Apple leaf disease detection method based on improved Yolov5s model |
| CN115345321A (en) * | 2022-10-19 | 2022-11-15 | 小米汽车科技有限公司 | Data augmentation method, data augmentation device, electronic device, and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107909638B (en) | Rendering method, medium, system and electronic device of virtual object | |
| CN110021052B (en) | Method and apparatus for generating fundus image generation model | |
| CN114444653A (en) | Method and system for evaluating influence of data augmentation on deep learning model performance | |
| EP3598386A1 (en) | Method and apparatus for processing image | |
| CN111460876B (en) | Method and apparatus for identifying video | |
| CN111325704B (en) | Image restoration method and device, electronic equipment and computer-readable storage medium | |
| CN112381717A (en) | Image processing method, model training method, device, medium, and apparatus | |
| CN114898177B (en) | Defect image generation method, model training method, device, medium and product | |
| US10572976B2 (en) | Enhancing observation resolution using continuous learning | |
| CN115905622A (en) | Video annotation method, device, equipment, medium and product | |
| CN112419179A (en) | Method, device, equipment and computer readable medium for repairing image | |
| CN111382647A (en) | Picture processing method, device, equipment and storage medium | |
| US11501530B1 (en) | Spatio-temporal relation based MR content positioning | |
| CN113411550B (en) | Video coloring method, device, equipment and storage medium | |
| CN111369475B (en) | Method and apparatus for processing video | |
| CN117171573A (en) | Training method, device, equipment and storage medium for multi-modal model | |
| Wang et al. | A multi-scale attentive recurrent network for image dehazing | |
| US20240177409A1 (en) | Image processing method and apparatus, electronic device, and readable storage medium | |
| CN113706663B (en) | Image generation method, device, equipment and storage medium | |
| CN112070888B (en) | Image generation method, device, equipment and computer readable medium | |
| CN117437516A (en) | Semantic segmentation model training method and device, electronic equipment and storage medium | |
| CN114066722A (en) | Method and device for acquiring image and electronic equipment | |
| CN113177176A (en) | Feature construction method, content display method and related device | |
| CN113762260A (en) | Method, device and equipment for processing layout picture and storage medium | |
| CN110599437A (en) | Method and apparatus for processing video |
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 |