Positive sample acquisition method and device for target detection, equipment and storage medium
A technology of target detection and acquisition method, applied in the field of image recognition, it can solve the problems of reducing algorithm performance, introducing noise in positive samples, and the model cannot return to an object autonomously, so as to solve the problem of insufficient training samples and low quality and high performance. Effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Benefits of technology
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0034] figure 1 It is a flowchart of a positive sample acquisition method for target detection provided by Embodiment 1 of the present invention. This embodiment is applicable to the case of screening samples for training models in image recognition, and specifically includes the following steps:
[0035] Step 1000, acquire an image to be detected, where the image to be detected includes at least one person image.
[0036] In this embodiment, the image to be detected is an image captured by a camera device for image recognition, which includes a person image or multiple person images. In an alternative embodiment, in addition to the person image, the image to be detected may also be an animal image, a vehicle image or other camera material images.
[0037] Step 1100, obtaining the anchor frame of the image to be detected.
[0038] In this embodiment, the target detection algorithm usually samples a large number of areas in the input image, and then judges whether these areas co
Embodiment 2
[0052] Please refer to figure 2 , figure 2 It is a flow chart of a method for obtaining positive samples for target detection provided by Embodiment 2 of the present invention. This embodiment optimizes and improves based on the solution of the first embodiment, and in particular provides a specific solution for matching rules between the anchor frame and the real label frame. Such as figure 2 As shown, the method includes:
[0053] Step 2000, acquire an image to be detected, and the image to be detected includes at least one person image.
[0054] Step 2100, acquire the anchor frame of the image to be detected.
[0055] Step 2200. Obtain the real annotation frame of the person image.
[0056] Similarly, step S2100 and step S2200 do not have a certain logical sequence relationship, step S2200 may be executed first and then step S2100 may be executed, or step S2100 and step S2200 may be executed simultaneously.
[0057] Step 2300, matching all the ground truth boxes with
Embodiment 3
[0070] Please refer to image 3 , image 3 It is a flow chart of a positive sample acquisition method for target detection provided by Embodiment 3 of the present invention. This embodiment optimizes and improves based on the solution of the first embodiment, and in particular provides a specific solution for filtering the first matching pair set according to the preset screening rules to obtain the second matching pair set. Such as image 3 As shown, the method includes:
[0071] Step 3000, acquire an image to be detected, where the image to be detected includes at least one person image.
[0072] Step 3100, acquire the anchor frame of the image to be detected.
[0073] Step 3200. Obtain the real annotation frame of the person image.
[0074] Similarly, step S3100 and step S3200 do not have a certain logical sequence relationship, step S3200 may be executed first and then step S3100 may be executed, or step S3100 and step S3200 may be executed simultaneously.
[0075] Step
PUM
Abstract
Description
Claims
Application Information
- 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