CN112906789A - Method and device for selecting negative examples of area generation network and computer equipment - Google Patents
Method and device for selecting negative examples of area generation network and computer equipment Download PDFInfo
- Publication number
- CN112906789A CN112906789A CN202110191775.1A CN202110191775A CN112906789A CN 112906789 A CN112906789 A CN 112906789A CN 202110191775 A CN202110191775 A CN 202110191775A CN 112906789 A CN112906789 A CN 112906789A
- Authority
- CN
- China
- Prior art keywords
- negative
- negative example
- positive
- candidate
- sample set
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a method, a device and computer equipment for selecting a negative example of an area generation network, wherein the method for selecting the negative example of the area generation network comprises the following steps: after a positive example sample set of the area generation network is obtained, obtaining a plurality of negative example candidate frames which are consistent with the size parameters of all positive examples in the positive example sample set and serve as the negative example candidate frame set; calculating the quantity proportion among different size parameters in the normal sample set; and screening out negative example training samples with corresponding proportions from the negative example candidate frame set according to the quantity proportion to serve as a negative example sample set. According to the area generation network negative example selection method, the negative example samples which are consistent with the size parameters of the positive example samples are screened out, the negative example samples which are consistent with the quantity proportion of the size parameters of the positive example sample set are formed, the strength of subsequent network negative example sample training can be increased, the correlation between the negative example sample training and the positive example sample training is enhanced, the accuracy of image detection classification is effectively improved, and false detection is reduced.
Description
Technical Field
The invention relates to the field of image recognition, in particular to a method and a device for selecting a negative case of a regional generation network, computer equipment and a readable storage medium.
Background
The existing image detection network is generally divided into two stages of networks, and comprises an area generation network and a subsequent classification detection network, negative examples of training samples are generally selected in a random mode from candidate frame training samples obtained through the area generation network, so that the negative examples of the subsequent two networks cannot be sufficiently trained, the training aid for the positive examples is small, and when the features of a target image are not obvious, the image detection network trained through the random negative examples is low in accuracy and high in false detection rate.
Disclosure of Invention
In view of the above problems, the present invention provides a method, an apparatus, a computer device and a readable storage medium for selecting a negative case of an area generation network, so as to effectively improve the accuracy of image detection and classification and reduce false detection.
In order to achieve the purpose, the invention adopts the following technical scheme:
a negative case selection method for an area generation network comprises the following steps:
after a positive example sample set of the area generation network is obtained, obtaining a plurality of negative example candidate frames which are consistent with the size parameters of all positive examples in the positive example sample set and serve as the negative example candidate frame set;
calculating the quantity proportion among different size parameters in the normal sample set;
and screening out negative example training samples with corresponding proportions from the negative example candidate frame set according to the quantity proportion to serve as a negative example sample set.
Preferably, in the method for selecting a negative example of an area-generated network, the obtaining a plurality of negative example candidate frames that are consistent with the size parameters of all positive examples in the positive example sample set includes:
after the current due sample is determined, recording the size parameter of the current due sample into a size parameter set;
and when obtaining negative example samples, screening out a preset number of negative example samples of various size parameters in the size parameter set from the negative example samples, and generating the negative example candidate frame set.
Preferably, the method for selecting the negative case of the area generation network further includes:
and carrying out negative example training on the image classification detection network by using the negative example sample set.
Preferably, in the negative selection method for the area-generating network, the size parameter includes a candidate box length, a candidate box width, and/or a candidate box area.
The invention also provides a device for selecting the negative case of the area generation network, which comprises the following components:
the candidate frame set acquisition module is used for acquiring a plurality of negative example candidate frames which are consistent with the size parameters of all positive examples in the positive example sample set and serve as a negative example candidate frame set after the positive example sample set of the area generation network is acquired;
the quantity proportion calculation module is used for calculating the quantity proportion among different size parameters in the normal sample set;
and the negative example sample acquisition module is used for screening out negative example training samples with corresponding proportions from the negative example candidate frame set according to the quantity proportion to serve as a negative example sample set.
Preferably, in the negative selection apparatus for an area-generated network, the candidate frame set obtaining module includes:
the size parameter recording unit is used for recording the size parameter of the current due sample into the size parameter set after the current due sample is determined;
and the candidate frame set generating unit is used for screening out a preset number of negative examples of various size parameters in the size parameter set from the negative examples when the negative examples are obtained, and generating the negative example candidate frame set.
Preferably, the negative case selecting device for an area generation network further includes:
and the negative case training module is used for carrying out negative case training on the image classification detection network by using the negative case sample set.
Preferably, in the negative example selecting apparatus for area-generated networks, the size parameter includes a frame candidate length, a frame candidate width, and/or a frame candidate area.
The invention also provides a computer device, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor runs the computer program to enable the computer device to execute the negative example selection method of the area generation network.
The invention also provides a readable storage medium storing a computer program which, when run on a processor, performs the method for area-generating network negative case selection.
The invention provides a selection method of a negative case of an area generation network, which comprises the following steps: after a positive example sample set of the area generation network is obtained, obtaining a plurality of negative example candidate frames which are consistent with the size parameters of all positive examples in the positive example sample set and serve as the negative example candidate frame set; calculating the quantity proportion among different size parameters in the normal sample set; and screening out negative example training samples with corresponding proportions from the negative example candidate frame set according to the quantity proportion to serve as a negative example sample set. According to the area generation network negative example selection method, the negative example samples which are consistent with the size parameters of the positive example samples are screened out, the negative example samples which are consistent with the quantity proportion of the size parameters of the positive example sample set are formed, the strength of subsequent network negative example sample training can be increased, the correlation between the negative example sample training and the positive example sample training is enhanced, the accuracy of image detection classification is effectively improved, and false detection is reduced.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings required to be used in the embodiments will be briefly described below, and it should be understood that the following drawings only illustrate some embodiments of the present invention, and therefore should not be considered as limiting the scope of the present invention. Like components are numbered similarly in the various figures.
Fig. 1 is a flowchart of a negative example selection method for an area generation network according to embodiment 1 of the present invention;
FIG. 2 is a flowchart of generating a negative example candidate box set according to embodiment 2 of the present invention;
fig. 3 is a flowchart of a negative example selection method for an area generation network according to embodiment 3 of the present invention;
fig. 4 is a schematic structural diagram of a negative example selecting apparatus for an area generation network according to embodiment 4 of the present invention;
fig. 5 is a schematic structural diagram of a candidate frame set obtaining module according to embodiment 4 of the present invention;
fig. 6 is a schematic structural diagram of another negative example selection apparatus for an area generation network according to embodiment 4 of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments.
The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.
Hereinafter, the terms "including", "having", and their derivatives, which may be used in various embodiments of the present invention, are only intended to indicate specific features, numbers, steps, operations, elements, components, or combinations of the foregoing, and should not be construed as first excluding the existence of, or adding to, one or more other features, numbers, steps, operations, elements, components, or combinations of the foregoing.
Furthermore, the terms "first," "second," "third," and the like are used solely to distinguish one from another and are not to be construed as indicating or implying relative importance.
Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which various embodiments of the present invention belong. The terms (such as those defined in commonly used dictionaries) should be interpreted as having a meaning that is consistent with their contextual meaning in the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein in various embodiments of the present invention.
Example 1
Fig. 1 is a flowchart of a negative-case selection method for an area-generated network according to embodiment 1 of the present invention, where the method includes the following steps:
step S11: after a positive example sample set of the area generation network is acquired, a plurality of negative example candidate frames which are consistent with the size parameters of all positive examples in the positive example sample set are acquired as a negative example candidate frame set.
In the embodiment of the invention, the image detection classification algorithm based on deep learning is generally divided into two parts, including an area generation network and a classification detection network. The area generation network is used for acquiring candidate frames in the image, that is, acquiring an image area which needs to be subjected to image classification detection. When the area generation network is used for sample training, training is generally performed through positive example samples and negative example samples, for example, a title in a document picture needs to be identified through image detection, the positive example samples when the area generation network is generated in the training area may be candidate frames whose intersection with a title area is greater than 0.7, and the candidate frames whose intersection is less than 0.7 may be negative example samples.
In the embodiment of the present invention, before training the area generation network, a training picture may be input to the area generation network to obtain a large number of random candidate boxes, so as to select suitable positive samples and negative samples from the candidate boxes for subsequent training. For positive example samples, candidate frames whose intersection with the target identification region is greater than a preset value may be screened from all the generated candidate frames, and the preset value may be 0.7, 0.8, and the like, which is not limited herein. The obtained plurality of positive example samples can generate a positive example sample set for use in generating a network and classifying a detection network in a subsequent training area.
In the embodiment of the present invention, when a negative example candidate frame is selected from the candidate frames, a plurality of negative example candidate frames are obtained according to the size parameter of each positive example in the positive example sample set, for example, the positive example sample set includes two positive example samples with size parameters of 1 and 2, and when a negative example candidate frame is selected, two negative example candidate frames with size parameters of 1 and 2 are also selected instead of being randomly selected. Also, when the negative example candidate frame is preliminarily selected, a preset number may be selected, and the preset number may be set by a user, or all negative example candidate frames that meet the size parameter may be selected, which is not limited herein.
In an embodiment of the present invention, the size parameter includes a frame candidate length, a frame candidate width, and/or a frame candidate area. The size parameter determination for the candidate box of the positive example or the negative example can be implemented by an algorithm or an application program, for example, an application program for determining the size parameter can be provided in the computer device, and the candidate box can be input to the application program after being generated to obtain the corresponding size parameter.
Step S12: and calculating the quantity proportion among different size parameters in the normal sample set.
In the embodiment of the invention, in order to make the training of the corresponding negative example sample and the training of the corresponding positive example sample reach the same strength in the training process, the quantitative ratio between different sizes of the negative example sample should correspond to the positive example sample. Therefore, before the final screening of the negative examples, the quantity ratio between the different size parameters in the positive example set needs to be calculated, for example, if the quantity of the positive examples with the size parameter of 1 is 10 and the quantity of the positive examples with the size parameter of 2 is 100 in one positive example set, the quantity ratio between the size parameters 1 and 2 can be calculated to be 1: 10.
In the embodiment of the present invention, the process of calculating the quantity ratio may be implemented by using an algorithm or an application program, for example, a statistical application program may be set in the computer device, the quantity of the regular samples of various size parameters is obtained by the statistical application program, and finally the corresponding quantity ratio is calculated.
Step S13: and screening out negative example training samples with corresponding proportions from the negative example candidate frame set according to the quantity proportion to serve as a negative example sample set.
In this embodiment of the present invention, after the quantity proportion of the various size parameters in the positive example sample set is calculated, the negative example training samples with the corresponding proportion may be screened from the negative example candidate frame set according to the quantity proportion, for example, if the quantity proportion between the size parameters 1 and 2 is 1:10, the number of the screened negative example training samples with the size parameter 1 may be 100, and the size parameter 2 may be 1000, or the number of the negative example training samples with the size parameter 1 may be 50, and the size parameter 2 may be 500, which is not limited herein. The screening process may be implemented by using an algorithm or an application program, which is not limited herein.
In the embodiment of the invention, the negative example samples which are consistent with the size parameters of the positive example samples are screened out, and the negative example samples which are consistent with the quantity proportion of the size parameters of the positive example sample set are formed, so that the strength of the subsequent network for carrying out the training of the negative example samples can be increased, the correlation between the training of the negative example samples and the training of the positive example samples is enhanced, the accuracy of image detection classification is effectively improved, and false detection is reduced.
Example 2
Fig. 2 is a flowchart of generating a negative example candidate frame set according to embodiment 2 of the present invention, including the following steps:
step S21: after the current due sample is determined, the size parameter of the current due sample is recorded into the size parameter set.
In the embodiment of the present invention, an application program for recording size parameters may be set in the computer device, and after determining the current due sample from the generated candidate frame, the application program may record corresponding size parameters, and finally record the size parameters of different due samples as a size parameter set.
Step S22: and when obtaining negative example samples, screening out a preset number of negative example samples of various size parameters in the size parameter set from the negative example samples, and generating the negative example candidate frame set.
In the embodiment of the invention, when the negative sample is obtained from the candidate frame, one of the size parameters can be obtained from the size parameter set, a preset number of negative samples are obtained according to the current size parameter, and by analogy, a preset number of negative samples corresponding to all the size parameters in the size parameter set are obtained, and the negative candidate frame set is generated.
Example 3
Fig. 3 is a flowchart of a negative-case selection method for an area-generated network according to embodiment 3 of the present invention, where the method includes the following steps:
step S31: after a positive example sample set of the area generation network is acquired, a plurality of negative example candidate frames which are consistent with the size parameters of all positive examples in the positive example sample set are acquired as a negative example candidate frame set.
This step is identical to step S11 described above, and will not be described herein again.
Step S32: and calculating the quantity proportion among different size parameters in the normal sample set.
This step is identical to step S12 described above, and will not be described herein again.
Step S33: and screening out negative example training samples with corresponding proportions from the negative example candidate frame set according to the quantity proportion to serve as a negative example sample set.
This step is identical to step S13 described above, and will not be described herein again.
Step S34: and carrying out negative example training on the image classification detection network by using the negative example sample set.
In the embodiment of the invention, by enhancing the correlation between the negative example sample training and the positive example sample training, after the negative example sample training of the image classification detection network is carried out, the image classification accuracy of the positive example can be effectively improved, and particularly when the image target characteristics are not obvious, the false detection can be effectively reduced, and the user experience is improved.
Example 4
Fig. 4 is a schematic structural diagram of a negative example selection apparatus for an area generation network according to embodiment 4 of the present invention.
The area generation network negative example selection device 400 includes:
a candidate frame set obtaining module 410, configured to obtain, after obtaining a positive example sample set of the area generation network, multiple negative example candidate frames that are consistent with size parameters of all positive examples in the positive example sample set, as a negative example candidate frame set;
a quantity ratio calculation module 420, configured to calculate a quantity ratio between different size parameters in the normal sample set;
and a negative example sample obtaining module 430, configured to filter out negative example training samples of corresponding proportions from the negative example candidate frame set according to the quantity proportion, where the negative example training samples serve as a negative example sample set.
As shown in fig. 5, the candidate box set obtaining module 410 includes:
a size parameter recording unit 411, configured to record the size parameter of the current due sample into the size parameter set after determining the current due sample;
and the candidate frame set generating unit 412 is configured to, when negative examples are obtained, screen out a preset number of negative examples of various size parameters in the size parameter set from the negative examples, and generate the negative example candidate frame set.
As shown in fig. 6, the negative example selecting apparatus 400 for an area generation network further includes:
and the negative example training module 440 is configured to perform negative example training on the image classification detection network by using the negative example sample set.
In an embodiment of the present invention, the size parameter includes a candidate box length, a candidate box width, and/or a candidate box area.
In the embodiment of the present invention, for more detailed description of functions of the modules, reference may be made to contents of corresponding parts in the foregoing embodiment, which are not described herein again.
In addition, the invention also provides computer equipment which can comprise a smart phone, a tablet computer, a vehicle-mounted computer, intelligent wearable equipment and the like. The computer device comprises a memory and a processor, wherein the memory can be used for storing a computer program, and the processor executes the computer program, so that the computer device executes the functions of each module in the method or the area generation network negative example selection device.
The memory may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data (such as audio data, a phonebook, etc.) created according to the use of the computer device, and the like. Further, the memory may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device.
The present embodiment also provides a readable storage medium for storing a computer program used in the computer device described above.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative and, for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present invention. 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 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.
In addition, each functional module or unit in each embodiment of the present invention may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention or a part of the technical solution that contributes to the prior art in essence can be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a smart phone, a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A negative case selection method for a region-generated network is characterized by comprising the following steps:
after a positive example sample set of the area generation network is obtained, obtaining a plurality of negative example candidate frames which are consistent with the size parameters of all positive examples in the positive example sample set and serve as the negative example candidate frame set;
calculating the quantity proportion among different size parameters in the normal sample set;
and screening out negative example training samples with corresponding proportions from the negative example candidate frame set according to the quantity proportion to serve as a negative example sample set.
2. The method according to claim 1, wherein the obtaining a plurality of negative case candidate boxes corresponding to the size parameters of all positive cases in the positive case sample set comprises:
after the current due sample is determined, recording the size parameter of the current due sample into a size parameter set;
and when obtaining negative example samples, screening out a preset number of negative example samples of various size parameters in the size parameter set from the negative example samples, and generating the negative example candidate frame set.
3. The method of claim 1, further comprising:
and carrying out negative example training on the image classification detection network by using the negative example sample set.
4. The method according to claim 1, wherein the size parameter comprises a candidate box length, a candidate box width, and/or a candidate box area.
5. An area generation network negative case selection apparatus, comprising:
the candidate frame set acquisition module is used for acquiring a plurality of negative example candidate frames which are consistent with the size parameters of all positive examples in the positive example sample set and serve as a negative example candidate frame set after the positive example sample set of the area generation network is acquired;
the quantity proportion calculation module is used for calculating the quantity proportion among different size parameters in the normal sample set;
and the negative example sample acquisition module is used for screening out negative example training samples with corresponding proportions from the negative example candidate frame set according to the quantity proportion to serve as a negative example sample set.
6. The apparatus according to claim 5, wherein the candidate box set obtaining module comprises:
the size parameter recording unit is used for recording the size parameter of the current due sample into the size parameter set after the current due sample is determined;
and the candidate frame set generating unit is used for screening out a preset number of negative examples of various size parameters in the size parameter set from the negative examples when the negative examples are obtained, and generating the negative example candidate frame set.
7. The negative selection device of an area generation network according to claim 5, further comprising:
and the negative case training module is used for carrying out negative case training on the image classification detection network by using the negative case sample set.
8. The apparatus according to claim 5, wherein the size parameter comprises a candidate box length, a candidate box width and/or a candidate box area.
9. A computer device comprising a memory storing a computer program and a processor executing the computer program to cause the computer device to perform the area generation network negative example selection method according to any one of claims 1 to 4.
10. A readable storage medium, characterized in that it stores a computer program which, when run on a processor, performs the area generation network negative case selection method of any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110191775.1A CN112906789A (en) | 2021-02-19 | 2021-02-19 | Method and device for selecting negative examples of area generation network and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110191775.1A CN112906789A (en) | 2021-02-19 | 2021-02-19 | Method and device for selecting negative examples of area generation network and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112906789A true CN112906789A (en) | 2021-06-04 |
Family
ID=76123841
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110191775.1A Pending CN112906789A (en) | 2021-02-19 | 2021-02-19 | Method and device for selecting negative examples of area generation network and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112906789A (en) |
-
2021
- 2021-02-19 CN CN202110191775.1A patent/CN112906789A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109544166B (en) | Risk identification method and risk identification device | |
| US7171042B2 (en) | System and method for classification of images and videos | |
| JP6240199B2 (en) | Method and apparatus for identifying object in image | |
| CN110781960B (en) | Training method, classification method, device and equipment of video classification model | |
| CN111881289B (en) | Training method of classification model, and detection method and device of data risk class | |
| KR102042168B1 (en) | Methods and apparatuses for generating text to video based on time series adversarial neural network | |
| CN111339250B (en) | Mining method for new category labels, electronic equipment and computer readable medium | |
| US11538141B2 (en) | Method and apparatus for processing video | |
| CN110738611A (en) | video image quality enhancement method, system and equipment | |
| CN106598997B (en) | Method and device for calculating text theme attribution degree | |
| CN117409419A (en) | Image detection method, device and storage medium | |
| CN110942034A (en) | Method, system and device for detecting multi-type depth network generated image | |
| CN112906789A (en) | Method and device for selecting negative examples of area generation network and computer equipment | |
| CN111062385A (en) | Network model construction method and system for image text information detection | |
| CN111757182B (en) | Image splash screen detection method, device, computer device and readable storage medium | |
| CN111724810A (en) | Audio classification method and device | |
| US20150288973A1 (en) | Method and device for searching for image | |
| CN114842382A (en) | Method, device, equipment and medium for generating semantic vector of video | |
| CN111078877B (en) | Data processing method, training method of text classification model, and text classification method and device | |
| CN111353428B (en) | Action information identification method and device, electronic equipment and storage medium | |
| KR102504321B1 (en) | Apparatus and method for online action detection | |
| CN113887674A (en) | Abnormal behavior detection method and system based on big data | |
| CN111310176B (en) | Intrusion detection method and device based on feature selection | |
| CN114241361A (en) | Video gene extraction and video matching method and device based on video gene | |
| CN114494887A (en) | Remote sensing image classification method and device, computer equipment and readable storage medium |
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 |