一种目标重识别方法、电子设备及存储介质
By calculating keypoint scores and pose confidence in video images, high-quality image sequences are selected for target re-identification, solving the problem of low recognition accuracy in complex scenes and achieving higher recognition accuracy.
CN116109846BActive Publication Date: 2026-07-17ZHEJIANG DAHUA TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ZHEJIANG DAHUA TECH CO LTD
- Filing Date
- 2022-12-30
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
In the process of video target recognition, due to the complexity and variability of the scene under video surveillance, problems such as dense crowds, pedestrian occlusion, changes in posture, and spatial misalignment are prone to occur, resulting in low recognition accuracy.
Method used
By acquiring multiple frames of images to be identified, calculating the key point score and pose confidence of each frame, and selecting image sequences with high integrity scores and multi-pose scores for target re-identification.
Benefits of technology
It improves the accuracy of target recognition in complex scenes, enhances image quality, and improves recognition performance.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN116109846B_ABST
Abstract
本发明公开了一种目标重识别方法、电子设备及存储介质，其中，方法包括：获取多帧待识别图像；获取每一帧待识别图像中的待识别目标对应的目标框，以及获取待识别目标对应的关键点评分和姿态置信度；其中，关键点评分是对待识别目标进行关键点识别得到；利用目标框和关键点评分的相关性，得到每一帧待识别图像的完整性分数；利用姿态置信度对多帧待识别图像进行筛选，得到筛选后的待识别图像序列，并确定待识别图像序列对应的多姿态分数；对多姿态分数和完整性分数满足预设条件的待识别图像序列进行目标重识别，得到目标识别结果。即本申请通过对图像的完整性分数和图像序列的多姿态分数，对图像进行优选，以提高对图像中目标的识别准确率。
Need to check novelty before this filing date? Find Prior Art