Unsupervised video clip retrieval method based on time sequence anchor point mining and semantic alignment

By using a method based on temporal anchor mining and semantic alignment, high-confidence temporal anchors are generated and point-supervised learning is performed. This solves the problems of false boundary noise and semantic alignment distortion in unsupervised video clip retrieval, and improves the accuracy and robustness of video clip retrieval.

CN122019834APending Publication Date: 2026-05-12SOUTH CHINA UNIV OF TECH
3 Cites 0 Cited by

Patent Information

Application Number
CN202610177907.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-02-07
Publication Date
2026-05-12

AI Technical Summary

Technical Problem

Existing unsupervised video clip retrieval methods suffer from severe false boundary noise, semantic alignment distortion, computational redundancy in the inference stage, and lack of robustness.

Method used

By constructing a method based on temporal anchor mining and semantic alignment, a key anchor extraction module is used to generate high-confidence temporal anchors. Combined with a semantic alignment description generation module and a point-supervised learning enhancement module, pseudo-label triples are constructed and point-supervised comparative learning is performed to improve the model's boundary awareness capability.

Benefits of technology

The model achieves both global semantic matching and accurate local temporal localization capabilities without the need for manual annotation, thus improving the localization accuracy and generalization ability of video segment retrieval.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses an unsupervised video clip retrieval method based on time sequence anchor point mining and semantic alignment, which comprises a video retrieval training system based on time sequence anchor point mining and point supervised learning, and the system comprises a key anchor point extraction module, a semantic alignment description generation module and a point supervised learning enhancement module. Key time anchor points are extracted from an original unlabeled video sequence through a key anchor point extraction module, then a pseudo-label triple is constructed through a semantic alignment description generation module, a weak supervision video clip retrieval model is trained accordingly, and in the training process, the pseudo-label triple is extracted from the original unlabeled video sequence. Constructing a point supervised contrast learning target through a point supervised learning enhancement module to further optimize the model, so that the finally trained weak supervised video clip retrieval model outputs a clip starting and ending time boundary related to query statement semantics; the problems of low pseudo tag quality, fuzzy positioning boundary and inconsistent semantic space in an unsupervised scene are effectively solved, and the positioning precision and generalization ability of video clip retrieval are remarkably improved.
Need to check novelty before this filing date? Find Prior Art