Weakly-supervised object detection method based on pseudo-true value search method
A technology of object detection and weak supervision, applied in image data processing, instrumentation, computing, etc., can solve the problems of inaccurate object position detection and inconsistency of categories.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
specific Embodiment approach 1
[0023] Specific implementation mode 1: The weakly supervised object detection method based on the false truth search method of this implementation mode includes:
[0024] Step 1), constructing training samples;
[0025] Step 2), input the picture in the training sample into the weakly supervised object detector based on multiple-instance learning method (Multiple-Instance Learning);
[0026] Step 3), performing non-maximum suppression processing on the output result of the weakly supervised object detector, retaining bounding boxes exceeding a predetermined score threshold in the processing results, and removing bounding boxes lower than the score threshold;
[0027] Step 4), among the bounding boxes retained in step 3), delete bounding boxes that are completely contained in other bounding boxes;
[0028] Step 5), for each bounding box obtained in step 4), calculate the overlapping area of the bounding box and other bounding boxes, and fuse the bounding boxes whose overlapp...
specific Embodiment approach 2
[0036] Specific implementation mode two: the difference between this implementation mode and specific implementation mode one is that step one specifically includes:
[0037] Step 1.1), receiving keywords input by the user; the keywords are used to indicate the category of the object;
[0038] Step 1.2), use the keyword to search in the search engine, select a preset number of search results and use the keyword as the annotation information of the search result.
[0039] That is, the present invention only needs to know the simple object category information in the picture, and can train the model without complex object position information. The simple object category information here can be obtained in many ways, such as searching for pictures in the search engine in the form of keywords ("pedestrians", "vehicles", etc.), and downloading the top few thousand pictures can be used as training samples , no manual labeling is required.
[0040] It can be understood that, when u...
specific Embodiment approach 3
[0042] Specific embodiment three: the difference between this embodiment and specific embodiment one or two is: in step 1), the training sample set can be any one of PASCAL VOC 2007 / 2012, MC COCO, WIDER FACE and FDDB database, or is The database constructed according to the method of the second specific embodiment.
[0043] Other steps and parameters are the same as those in Embodiment 1 or Embodiment 2.
PUM
![No PUM](https://static-eureka.patsnap.com/ssr/23.2.0/_nuxt/noPUMSmall.5c5f49c7.png)
Abstract
Description
Claims
Application Information
![application no application](https://static-eureka.patsnap.com/ssr/23.2.0/_nuxt/application.06fe782c.png)
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap