Sports video highlight fragment intelligent extraction method and system for ball competition

By constructing a unified temporal state representation and a specular potential function, and combining multimodal features and game state features, the problem of inaccurate boundaries and repetition in specular clip extraction from ball game videos was solved, achieving accurate, complete and efficient extraction of specular clips.

CN122454490APending Publication Date: 2026-07-24SHENZHEN AURORA LEADING TECHNOLOGY CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202610913682.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-06-24
Publication Date
2026-07-24

AI Technical Summary

Technical Problem

Existing technologies for extracting highlight clips from ball game videos lack continuous modeling of the game process, resulting in inaccurate highlight clip boundaries, truncation of action processes, difficulty in distinguishing actual contributions from superficial highlights, and issues of semantic similarity or content duplication.

Method used

By constructing a unified temporal state representation, introducing a constrained specular potential function, and combining multimodal features and competition state features, specular candidate segments are generated. Segment scoring is performed through causal counterfactual contribution and action chain coupling function to achieve global non-redundant optimization and output high-value and low-repetition specular segments.

Benefits of technology

It improves the accuracy of highlight fragment boundaries, ensures the integrity of action chains and the actual contribution assessment of fragments, reduces the output of semantically similar or repetitive fragments, and enhances the information density and interpretability of highlight fragments.

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

Abstract

The invention discloses an intelligent extraction method and system for a highlight fragment of a sports video for a ball match, belongs to the technical field of video processing and sports analysis, and aims at the pain points of inaccurate boundary recognition, misjudgment of key fragments, truncation of action chains and redundant results in the prior art. Constructing a highlight potential energy function with match state constraints, and automatically identifying highlight candidate boundaries; in combination with causal anti-fact contribution calculation and trajectory coupling energy evaluation, multi-dimensional fragment scoring is realized; finally, through global non-redundancy optimization, a low-repetition highlight set is screened, the highlight boundary can be accurately positioned, surface wonderful and actual key segments can be distinguished, the action chain integrity is ensured, the repeated content is effectively removed, the accuracy of highlight extraction of the ball competition and the user experience are greatly improved, and the method is suitable for popularization and application. The method is suitable for automatic video processing of various ball events.
Need to check novelty before this filing date? Find Prior Art