CN114494730A - Trademark automatic classification processing system based on image recognition - Google Patents
Trademark automatic classification processing system based on image recognition Download PDFInfo
- Publication number
- CN114494730A CN114494730A CN202210393064.7A CN202210393064A CN114494730A CN 114494730 A CN114494730 A CN 114494730A CN 202210393064 A CN202210393064 A CN 202210393064A CN 114494730 A CN114494730 A CN 114494730A
- Authority
- CN
- China
- Prior art keywords
- image
- module
- processed
- processing system
- classification
- 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
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/55—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- 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)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Library & Information Science (AREA)
- Databases & Information Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image recognition, in particular to an automatic trademark classification processing system based on image recognition, which comprises an acquisition module, a classification module and a classification module, wherein the acquisition module is used for carrying out target extraction processing on an image in a main body area and taking the extracted image as an image to be processed; the image matching module is used for carrying out image matching processing on the image to be processed and an example image in a specified characteristic region; the prediction module predicts the visibility of the preset key points of the image to be processed according to a visibility prediction algorithm; the classification module is used for determining the position and visibility of the key point of the image to be processed according to a preset image classification algorithm, extracting a classification result corresponding to local feature splicing prediction, and the invention solves the problem that the existing trademark recognition system cannot automatically classify trademarks.
Description
Technical Field
The invention relates to the technical field of image recognition, in particular to an automatic trademark classification processing system based on image recognition.
Background
Trademarks play a significant role in the development of socioeconomic performance as an important intellectual property. How to quickly and efficiently retrieve an approximate trademark from a vast array of trademarks would be very helpful to trademark reviewers and to the applicant of the trademark owner. Trademarks are one of the important logos of commodities, and the registration of trademarks is important for protecting the legal rights of the trademark holders, so that the new trademarks must be searched in a trademark library before being registered so as to ensure that the new trademarks are obviously different from other registered trademarks. As the official patent library is digitalized and started late, the included trademark images are not complete. Therefore, the official trademark database has a problem of data unbalance as with other data sets. If the training data is unbalanced, most of the existing learning algorithms generate learning bias for a plurality of classes, so that the recognition performance of a few classes is poor. The image retrieval technology based on the content is that a machine automatically extracts visual characteristics including color, texture, shape, position and mutual relation of objects and the like of image content, similar matching is carried out on images in a database and query sample images in a characteristic space, and images similar to the samples are retrieved.
The current methods for alleviating the data imbalance problem can be divided into two categories: under-sampling and over-sampling. However, for the task of searching trademarks, it is imperative to ensure the authenticity of the data; through the operation of kneading and synthesizing, even if the model training is better, the practical significance is lacked. The character trademark is easy to manage and can be searched by a character annotation method. However, image trademarks or combination trademarks containing images are not easily described in a text manner, and if a classification coding manner is adopted, the disadvantages of time-consuming classification, incomplete description subjectivity and the like exist, so that the image retrieval technology based on contents is gradually emphasized by people. Therefore, in the trademark retrieval task, the constructed training algorithm is not to increase the imbalance of the trademark data as much as possible while optimizing the retrieval performance. The number of data pairs which are dissimilar between the existing trademark image data is far larger than the number of similar pairs, so that the method for comprehensively learning from the relationship between the trademark data pairs has certain limitation.
Disclosure of Invention
Aiming at the defects of the prior art, the invention discloses an automatic trademark classification processing system based on image recognition, which is used for solving the problem that the conventional trademark recognition system cannot automatically classify trademarks;
the invention is realized by the following technical scheme:
the invention discloses an automatic trademark classification processing system based on image recognition, which comprises:
the acquisition module is used for carrying out target extraction processing on the image in the main body area and taking the extracted image as an image to be processed;
the preprocessing module is used for performing contour tracing on the image to be processed through gray processing;
the image matching module is used for carrying out image matching processing on the image to be processed and an example image in a specified characteristic region;
the extraction module is used for carrying out contour detection on the image to be processed when the specified feature is confirmed to exist, verifying the detected region according to the parameterized apparent model corresponding to the specified feature and extracting the region to be identified corresponding to the specified feature;
the prediction module predicts the visibility of the preset key points of the image to be processed according to a visibility prediction algorithm;
and the classification module is used for determining the positions and visibility of the key points of the image to be processed according to a preset image classification algorithm and extracting a classification result corresponding to local feature splicing prediction.
Still further, the image matching module 3 is composed of the following sub-modules, including:
the image feature extraction module is used for determining a main body area containing a main body part of the image to be processed in the specified feature area according to preset main body area feature parameters, and the coordinates and the size of the specified feature area;
the similarity calculation module is used for calculating whether a region to be identified corresponding to the specified characteristic region exists in the image to be processed;
and the judging module is used for judging the image matching processing result of the image to be processed.
Furthermore, the automatic trademark classification processing system further comprises a segmentation module, wherein the segmentation module is used for performing superpixel segmentation processing on the image to be processed and determining an average value of colors of the processed superpixel blocks.
Furthermore, the trademark automatic classification processing system further comprises a background color module, and the background color module is used for determining the background color of the image to be processed according to the average value of the colors of the super pixel blocks.
Furthermore, the trademark automatic classification processing system further comprises a pure color judgment module, wherein the pure color judgment module is used for determining a pure color background score of the image to be processed, and judging whether the image to be processed is a pure color background according to the pure color background score, and the pure color background score is a ratio of the number of pixels contained in the super pixel block of the color broad value conforming to the background color to the total number of pixels of the image to be processed.
Further, the determining module is composed of the following sub-modules, including:
the detection submodule is used for detecting whether an area corresponding to the specified feature exists in the image to be processed;
and the confirming submodule is used for carrying out contour detection on the region corresponding to the specified feature in the image to be processed when the detection submodule confirms that the region corresponding to the specified feature exists, verifying the detected region according to the parameterized apparent model corresponding to the specified feature and confirming whether the region is a region to be identified corresponding to the specified feature.
Further, the confirming sub-module is configured to detect a region corresponding to the example image in the image to be processed according to a similarity matching detection algorithm, and verify the detected region based on the classification appearance model.
Furthermore, the automatic trademark classification processing system further comprises a saliency detection module, wherein the saliency detection module is used for performing saliency detection on the image to be processed when the characteristic region does not exist, and determining a main region of the image to be processed according to a saliency value of each pixel point in the detected image to be processed.
Furthermore, the trademark automatic classification processing system further comprises an improvement module, wherein the improvement module is used for performing superpixel segmentation processing on the image to be processed, determining average significance values of all pixels in each superpixel block after processing, sequentially judging whether the average significance value of each superpixel block is higher than a preset wide value, and updating the main body area when the average significance value of the superpixel block is higher than the set wide value.
Furthermore, after the classification module marks the classification category to which the image to be processed belongs, a classification label is set according to actual requirements to classify the image; the classification label is provided with corresponding global characteristics and local characteristic categories.
The invention has the beneficial effects that:
1. according to the method, the main body area of the image to be processed is obtained in different modes according to different conditions of whether the characteristic area corresponding to the specified characteristic exists in the image to be processed or not, the preset algorithm is executed in the image of the main body area, and the image corresponding to the algorithm is used as the main body part of the image to be processed. Therefore, on the basis of realizing automatic extraction of the image main body, the extracted main body is accurate and reliable, and the processing efficiency is improved.
2. According to the method, through historical data accumulation and conversion analysis of the short-term behavior of the user to the long-term behavior, characteristics which are not interesting to the user are more accurately analyzed and extracted, more accurate trademark classification basis is provided for subsequent recommendation, and content loss caused by classification errors is avoided to a greater extent; when the trademark features of the user are more accurate, the overall click rate of the recommended content is better improved.
3. The invention solves the problem that the current trademark image retrieval research is still basically in the experimental stage. The searching method overcomes the defects of the existing searching method in the aspects of scaling and rotation invariance, searching capability for geometric deformation, searching precision, consistency of images and human visual perception and the like, achieves a more effective searching mode, and integrates various algorithms to meet the searching requirement.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic diagram of a system structure of an automatic trademark classification processing system based on image recognition;
the reference numerals in the drawings denote: 1. an acquisition module; 2. a preprocessing module; 3. an image matching module; 4. an image feature extraction module; 5. a similarity calculation module; 6. a judgment module; 7. an extraction module; 8. a prediction module; 9. and (5) a classification module.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, 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 some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
The present embodiment provides an automatic trademark classification processing system based on image recognition, please refer to fig. 1, which includes:
the acquisition module 1 is used for performing target extraction processing on an image in the main body area and taking the extracted image as an image to be processed;
the preprocessing module 2 is used for performing contour tracing on the image to be processed through gray processing;
the image matching module 3 is used for carrying out image matching processing on the image to be processed and an example image in a specified characteristic region;
an extraction module 7, configured to perform contour detection on the image to be processed when the specified feature is determined to exist, verify a detected region according to a parameterized appearance model corresponding to the specified feature, and extract a region to be identified corresponding to the specified feature;
the prediction module 8 predicts the visibility of the preset key points of the image to be processed according to a visibility prediction algorithm;
and the classification module 9 is configured to determine the position and visibility of the key point of the image to be processed according to a preset image classification algorithm, and extract a classification result corresponding to local feature stitching prediction.
The image matching module 3 is composed of the following sub-modules, including:
the image feature extraction module 4 is configured to determine, in the designated feature region, a main body region including a main body portion of the image to be processed according to preset main body region feature parameters, and coordinates and a size of the designated feature region;
the similarity calculation module 5 is configured to calculate whether a to-be-identified region corresponding to the specified feature region exists in the to-be-processed image;
and the judging module 6 is used for judging the image matching processing result of the image to be processed.
The trademark automatic classification processing system also comprises a segmentation module, wherein the segmentation module is used for carrying out superpixel segmentation processing on the image to be processed and determining the average value of the color of each processed superpixel block; the trademark automatic classification processing system also comprises a background color module, wherein the background color module is used for determining the background color of the image to be processed according to the average value of the colors of the super pixel blocks; the trademark automatic classification processing system also comprises a pure color judgment module, wherein the pure color judgment module is used for determining the pure color background score of the image to be processed and judging whether the image to be processed is a pure color background or not according to the pure color background score, and the pure color background score is the ratio of the number of pixels contained in the super pixel block of the color broad value conforming to the background color to the total number of pixels of the image to be processed.
After the classification module 9 marks the classification category to which the image to be processed belongs, a classification label is set according to actual requirements to classify the image; the classification label is provided with corresponding global characteristics and local characteristic categories.
According to the method, the main body area of the image to be processed is obtained in different modes according to different conditions of whether the characteristic area corresponding to the specified characteristic exists in the image to be processed or not, the preset algorithm is executed in the image of the main body area, and the image corresponding to the algorithm is used as the main body part of the image to be processed. Therefore, on the basis of realizing automatic extraction of the image main body, the extracted main body is accurate and reliable, and the processing efficiency is improved.
According to the method, through historical data accumulation and conversion analysis of the short-term behavior of the user to the long-term behavior, characteristics which are not interesting to the user are more accurately analyzed and extracted, more accurate trademark classification basis is provided for subsequent recommendation, and content loss caused by classification errors is avoided to a greater extent; when the trademark features of the user are more accurate, the overall click rate of the recommended content is better improved.
The invention solves the problem that the current trademark image retrieval research is still basically in the experimental stage. The method overcomes the defects of the existing retrieval method in aspects of scaling and rotation invariance, retrieval capability to geometric deformation, retrieval precision, consistency of images and human visual perception and the like, achieves a more effective retrieval mode, and integrates various algorithms to meet the retrieval requirements.
Example 2
In a specific implementation aspect, on the basis of embodiment 1, this embodiment further specifically describes the power grid security real-time detection system in embodiment 1 with reference to fig. 1, where the determining module 6 is composed of the following sub-modules, and includes:
the detection submodule is used for detecting whether an area corresponding to the specified feature exists in the image to be processed;
and the confirming submodule is used for carrying out contour detection on the region corresponding to the specified feature in the image to be processed when the detection submodule confirms that the region corresponding to the specified feature exists, verifying the detected region according to the parameterized apparent model corresponding to the specified feature and confirming whether the region is a region to be identified corresponding to the specified feature.
And the confirmation submodule is used for detecting the area corresponding to the example image in the image to be processed according to a similarity matching detection algorithm and verifying the detected area based on the classification appearance model.
According to the method, the main body area of the image to be processed is obtained in different modes according to different conditions of whether the characteristic area corresponding to the specified characteristic exists in the image to be processed or not, the preset algorithm is executed in the image of the main body area, and the image corresponding to the algorithm is used as the main body part of the image to be processed. Therefore, on the basis of realizing automatic extraction of the image main body, the extracted main body is accurate and reliable, and the processing efficiency is improved.
According to the method, through historical data accumulation and conversion analysis of the short-term behavior of the user to the long-term behavior, characteristics which are not interesting to the user are more accurately analyzed and extracted, more accurate trademark classification basis is provided for subsequent recommendation, and content loss caused by classification errors is avoided to a greater extent; when the trademark features of the user are more accurate, the overall click rate of the recommended content is better improved.
The invention solves the problem that the current trademark image retrieval research is still basically in the experimental stage. The method overcomes the defects of the existing retrieval method in aspects of scaling and rotation invariance, retrieval capability to geometric deformation, retrieval precision, consistency of images and human visual perception and the like, achieves a more effective retrieval mode, and integrates various algorithms to meet the retrieval requirements.
Example 3
In a specific implementation level, on the basis of embodiment 2, this embodiment further specifically describes the power grid security real-time detection system in embodiment 2 with reference to fig. 1, where the trademark automatic classification processing system further includes a saliency detection module, where the saliency detection module is configured to perform saliency detection on the image to be processed when the feature region does not exist, and determine a main region of the image to be processed according to a saliency value of each pixel point in the detected image to be processed.
The trademark automatic classification processing system also comprises an improved module, a main body area updating module and a display module, wherein the improved module is used for carrying out superpixel segmentation processing on the image to be processed, determining the average significance values of all pixels in the superpixel blocks after processing, sequentially judging whether the average significance value of each superpixel block is higher than a preset wide value or not, and updating the main body area when the average significance value of each superpixel block is higher than the set wide value.
The saliency detection module is specifically configured to perform binarization processing on a saliency map formed by the saliency values, perform elimination processing on smaller color blocks in an image after the binarization processing, find contour lines of the color blocks in the image after the elimination processing, approximate the contour lines of the color blocks by using polygons, and surround each polygon by using a rectangle as the main region.
According to the method, the main body area of the image to be processed is obtained in different modes according to different conditions of whether the characteristic area corresponding to the specified characteristic exists in the image to be processed or not, the preset algorithm is executed in the image of the main body area, and the image corresponding to the algorithm is used as the main body part of the image to be processed. Therefore, on the basis of realizing automatic extraction of the image main body, the extracted main body is accurate and reliable, and the processing efficiency is improved.
According to the method, through historical data accumulation and conversion analysis of the short-term behavior of the user to the long-term behavior, characteristics which are not interesting to the user are more accurately analyzed and extracted, more accurate trademark classification basis is provided for subsequent recommendation, and content loss caused by classification errors is avoided to a greater extent; when the trademark features of the user are more accurate, the overall click rate of the recommended content is better improved.
The invention solves the problem that the current trademark image retrieval research is still in the experimental stage basically. The method overcomes the defects of the existing retrieval method in aspects of scaling and rotation invariance, retrieval capability to geometric deformation, retrieval precision, consistency of images and human visual perception and the like, achieves a more effective retrieval mode, and integrates various algorithms to meet the retrieval requirements.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. Trademark automatic classification processing system based on image recognition is characterized by comprising:
the device comprises an acquisition module (1) and a processing module, wherein the acquisition module is used for carrying out target extraction processing on an image in a main body area and taking the extracted image as an image to be processed;
the preprocessing module (2) is used for performing contour tracing on the image to be processed through gray processing;
the image matching module (3) is used for carrying out image matching processing on the image to be processed and an example image in a specified characteristic region;
an extraction module (7) for detecting the contour of the image to be processed when the specified feature is confirmed to exist, verifying the detected region according to the parameterized apparent model corresponding to the specified feature, and extracting the region to be identified corresponding to the specified feature;
the prediction module (8) predicts the visibility of the preset key points of the image to be processed according to a visibility prediction algorithm;
and the classification module (9) is used for determining the positions of the key points of the image to be processed and extracting the classification result corresponding to local feature splicing prediction according to a preset image classification algorithm.
2. The image recognition-based trademark automatic classification processing system according to claim 1, characterized in that the image matching module (3) is composed of the following sub-modules including:
the image feature extraction module (4) is used for determining a main body area containing a main body part of the image to be processed in the specified feature area according to preset main body area feature parameters, and the coordinates and the size of the specified feature area;
the similarity calculation module (5) is used for calculating whether a region to be identified corresponding to the specified characteristic region exists in the image to be processed;
and the judging module (6) is used for judging the image matching processing result of the image to be processed.
3. The trademark automatic classification processing system based on image recognition according to claim 1, further comprising a segmentation module, wherein the segmentation module is configured to perform superpixel segmentation processing on the image to be processed, and determine an average value of colors of processed superpixel blocks.
4. The trademark automatic classification processing system based on image recognition according to claim 3, characterized in that the trademark automatic classification processing system further comprises a background color module, wherein the background color module is used for determining the background color of the image to be processed according to the average value of the colors of the super pixel blocks.
5. The trademark automatic classification processing system based on image recognition according to claim 4, further comprising a pure color judgment module, wherein the pure color judgment module is configured to determine a pure color background score of the image to be processed, and judge whether the image to be processed is a pure color background according to the pure color background score, and the pure color background score is a ratio of a number of pixels included in a super pixel block with a color broad value according to the background color to a total number of pixels of the image to be processed.
6. The image recognition-based trademark automatic classification processing system according to claim 1, characterized in that: the judging module (6) is composed of the following sub-modules, and comprises the following components:
the detection submodule is used for detecting whether an area corresponding to the specified feature exists in the image to be processed;
and the confirming submodule is used for carrying out contour detection on the region corresponding to the specified feature in the image to be processed when the detection submodule confirms that the region corresponding to the specified feature exists, verifying the detected region according to the parameterized apparent model corresponding to the specified feature and confirming whether the detected region is the region to be identified corresponding to the specified feature.
7. The image recognition-based trademark automatic classification processing system according to claim 6, wherein: and the confirming sub-module is used for detecting the area corresponding to the example image in the image to be processed according to a similarity matching detection algorithm and verifying the detected area based on a classification appearance model.
8. The trademark automatic classification processing system based on image recognition according to claim 1, further comprising a saliency detection module, wherein the saliency detection module is configured to perform saliency detection on the image to be processed when the feature region does not exist, and determine a main region of the image to be processed according to a saliency value of each pixel point in the detected image to be processed.
9. The image-recognition-based trademark automatic classification processing system according to claim 8, wherein the trademark automatic classification processing system further comprises an improvement module, the improvement module is configured to perform superpixel segmentation processing on the image to be processed, determine average saliency values of all pixels in each superpixel block after processing, sequentially judge whether the average saliency value of each superpixel block is higher than a preset broad value, and update the main body area when the average saliency value of each superpixel block is higher than a set broad value.
10. The image recognition-based trademark automatic classification processing system according to claim 1, wherein: after the classification module (9) marks the classification category to which the image to be processed belongs, setting a classification label according to actual requirements to classify the image; the classification label is provided with corresponding global characteristics and local characteristic categories.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210393064.7A CN114494730A (en) | 2022-04-15 | 2022-04-15 | Trademark automatic classification processing system based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210393064.7A CN114494730A (en) | 2022-04-15 | 2022-04-15 | Trademark automatic classification processing system based on image recognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114494730A true CN114494730A (en) | 2022-05-13 |
Family
ID=81489506
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210393064.7A Pending CN114494730A (en) | 2022-04-15 | 2022-04-15 | Trademark automatic classification processing system based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114494730A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115100450A (en) * | 2022-08-26 | 2022-09-23 | 北京海上升科技有限公司 | Intelligent traffic brand automobile big data detection method and system based on artificial intelligence |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105631455A (en) * | 2014-10-27 | 2016-06-01 | 阿里巴巴集团控股有限公司 | Image main body extraction method and system |
| CN108038122A (en) * | 2017-11-03 | 2018-05-15 | 福建师范大学 | A kind of method of trademark image retrieval |
| CN111091160A (en) * | 2019-12-27 | 2020-05-01 | 北京蜜莱坞网络科技有限公司 | Image classification method |
| US20210073521A1 (en) * | 2019-09-10 | 2021-03-11 | Amarjot Singh | Continuously Evolving and Interactive Disguised Face Identification (DFI) with Facial Key Points using ScatterNet Hybrid Deep Learning (SHDL) Network |
-
2022
- 2022-04-15 CN CN202210393064.7A patent/CN114494730A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105631455A (en) * | 2014-10-27 | 2016-06-01 | 阿里巴巴集团控股有限公司 | Image main body extraction method and system |
| CN108038122A (en) * | 2017-11-03 | 2018-05-15 | 福建师范大学 | A kind of method of trademark image retrieval |
| US20210073521A1 (en) * | 2019-09-10 | 2021-03-11 | Amarjot Singh | Continuously Evolving and Interactive Disguised Face Identification (DFI) with Facial Key Points using ScatterNet Hybrid Deep Learning (SHDL) Network |
| CN111091160A (en) * | 2019-12-27 | 2020-05-01 | 北京蜜莱坞网络科技有限公司 | Image classification method |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115100450A (en) * | 2022-08-26 | 2022-09-23 | 北京海上升科技有限公司 | Intelligent traffic brand automobile big data detection method and system based on artificial intelligence |
| CN115100450B (en) * | 2022-08-26 | 2023-01-06 | 中咨数据有限公司 | Intelligent traffic brand automobile big data detection method and system based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shi et al. | Automatic road crack detection using random structured forests | |
| CN107833213B (en) | Weak supervision object detection method based on false-true value self-adaptive method | |
| CN107257161B (en) | Transformer substation disconnecting link remote control auxiliary checking method and system based on state recognition algorithm | |
| Xu | Robust traffic sign shape recognition using geometric matching | |
| CN105117740B (en) | Font identification method and apparatus | |
| CN109299307B (en) | Trademark retrieval early warning method and device based on structural analysis | |
| CN104881668B (en) | A kind of image fingerprint extracting method and system based on representative local mode | |
| WO2023035558A1 (en) | Anchor point cut-based image processing method and apparatus, device, and medium | |
| CN114494730A (en) | Trademark automatic classification processing system based on image recognition | |
| CN112926601A (en) | Image recognition method, device and equipment based on deep learning and storage medium | |
| CN114882010A (en) | Surface defect detection method based on picture recognition | |
| Karaoglu et al. | A novel algorithm for text detection and localization in natural scene images | |
| Muzakir et al. | Model for Identification and Prediction of Leaf Patterns: Preliminary Study for Improvement | |
| CN108428234B (en) | Interactive segmentation performance optimization method based on image segmentation result evaluation | |
| CN113723314A (en) | Sugarcane stem node identification method based on YOLOv3 algorithm | |
| CN110378337B (en) | Visual input method and system for drawing identification information of metal cutting tool | |
| CN116052110B (en) | Intelligent positioning method and system for pavement marking defects | |
| CN109857897B (en) | Trademark image retrieval method and device, computer equipment and storage medium | |
| CN111079752A (en) | Method and device for identifying circuit breaker in infrared image and readable storage medium | |
| CN114092542A (en) | Bolt measuring method and system based on two-dimensional vision | |
| CN113158743B (en) | Small target real-time detection and positioning method, system and equipment based on priori knowledge | |
| JP3514085B2 (en) | Character string input method and apparatus, and storage medium storing character string input program | |
| CN118071831A (en) | Image coarse positioning method, device and computer readable storage medium | |
| CN114037607A (en) | Fragment specimen splicing method, device, equipment and storage medium | |
| Huang et al. | A region and feature-based matching algorithm for dynamic object recognition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220513 |
|
| RJ01 | Rejection of invention patent application after publication |