Method and apparatus for determining counterfeit video on basis of multi-modal features, and device and medium
By extracting frame and audio features from videos, dynamically adjusting weights, and performing multi-level anomaly detection, the problem of low accuracy in detecting forged videos in existing technologies is solved, and efficient and accurate identification of complex forged videos is achieved.
WO2026129745A1PCT designated stage Publication Date: 2026-06-25CHINA TELECOM ARTIFICIAL INTELLIGENCE TECHNOLOGY (BEIJING) CO LTD
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- CHINA TELECOM ARTIFICIAL INTELLIGENCE TECHNOLOGY (BEIJING) CO LTD
- Filing Date
- 2025-09-05
- Publication Date
- 2026-06-25
Smart Images
Figure CN2025119425_25062026_PF_FP_ABST
Abstract
The present application relates to the technical field of artificial intelligence. Disclosed are a method and apparatus for determining a counterfeit video on the basis of multi-modal features, and a device and a medium. The method comprises: extracting picture frames and audio from a video; identifying objects from the picture frames, and then determining picture features of the objects; identifying audio features of the audio; on the basis of the picture features and the audio features, dynamically outputting a first weight, which relates to temporal synchronization between the picture frames and the audio, a second weight, which relates to spatio-temporal consistency between the picture frames and the audio, and a third weight, which relates to detecting whether features at different levels are abnormal, and extracting a time-series feature of the video; and on the basis of the first weight, the second weight, the third weight and the time-series feature, determining whether the video is counterfeit.
Need to check novelty before this filing date? Find Prior Art