Lowbrow image recognition method and device based on artificial intelligence and electronic equipment
By performing human body-level and body part-level classification processing on images, and combining human body and body part region matching, the problems of low accuracy and resource waste in vulgar image recognition are solved, achieving more efficient recognition and resource utilization.
CN115019336BActive Publication Date: 2026-06-19TENCENT TECHNOLOGY (SHENZHEN) CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TENCENT TECHNOLOGY (SHENZHEN) CO LTD
- Filing Date
- 2021-03-04
- Publication Date
- 2026-06-19
AI Technical Summary
Technical Problem
Existing technologies have low accuracy in identifying vulgar images, are prone to misjudgment, and have low utilization of computing resources.
Method used
By classifying images at the human body level and body part level, the matching of human body prediction regions and body part prediction regions is determined, and the vulgar image recognition result is determined by combining human body category and body part category.
Benefits of technology
It improves the accuracy of vulgar image recognition and enhances the actual utilization rate of computing resources.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN115019336B_ABST
Abstract
This application provides a method, apparatus, electronic device, and computer-readable storage medium for vulgar image recognition based on artificial intelligence. The method includes: performing human body-level classification processing on an image to obtain human body prediction regions and corresponding human body categories; performing body part-level classification processing on the image to obtain at least one body part prediction region and a body part category corresponding to each body part prediction region; performing position matching processing between the human body prediction regions and the at least one body part prediction region to obtain body part prediction regions that successfully match the human body prediction regions; and determining the vulgar image recognition result of the image based on the human body category corresponding to the human body prediction region and the body part category corresponding to the successfully matched body part prediction region. This application can improve the accuracy of vulgar image recognition and simultaneously improve the actual utilization rate of computing resources consumed by electronic devices in the vulgar image recognition process.
Need to check novelty before this filing date? Find Prior Art