Video anomaly detection method based on multi-task learning

By combining a spatiotemporal Transformer encoder, a prototype memory network, and a dual-decoder architecture, a multi-task learning method is used to solve the problems of Transformer overgeneralization and insufficient feature utilization in existing video anomaly detection, achieving higher accuracy and robustness in anomaly detection.

CN121982603APending Publication Date: 2026-05-05HANGZHOU HUISHI NUOBAO INTELLIGENT TECHNOLOGY CO LTD +1
0 Cites 0 Cited by

Patent Information

Application Number
CN202610059701.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-01-16
Publication Date
2026-05-05

AI Technical Summary

Technical Problem

Existing video anomaly detection methods suffer from problems such as overgeneralization of Transformer, task singularity, insufficient constraint mechanisms, and inadequate feature utilization, resulting in high false alarm rates and limited ability to detect complex anomaly types.

Method used

We employ a multi-task learning approach that combines a spatiotemporal Transformer encoder, a prototype memory network, and a dual-decoder architecture. By extracting hierarchical spatiotemporal features, storing normal pattern prototypes, and performing prediction and reconstruction tasks, we construct a self-supervised learning framework by leveraging the complementarity of original temporal features and memory-enhanced features.

Benefits of technology

It improves the accuracy and robustness of video anomaly detection, effectively constrains the model's generalization ability to anomaly samples, enhances the detection capability for different types of anomalies, and reduces the false alarm rate.

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

Abstract

The invention discloses a video anomaly detection method based on multi-task learning, a model comprises a space-time encoder, a prototype memory network, a double-decoder architecture and a reconstruction decoder, and the method comprises the following steps: carrying out hierarchical feature extraction on a training video frame sequence by using the space-time encoder; performing memory enhancement processing on the spatial-temporal characteristics through a prototype memory network; respectively carrying out future frame prediction and current frame reconstruction through a double-decoder architecture; calculating a multi-task loss function, and updating parameters of the space-time encoder, the prototype memory network and the double decoders through back propagation according to the multi-task loss function; according to the method, a more comprehensive self-supervised learning framework is constructed, and the generalization ability of the model is enhanced.
Need to check novelty before this filing date? Find Prior Art