2D bidirectional gated recurrent unit convolutional Neural networks for end-to-end violence detection In videos
This addresses the problem of automated violence detection for video surveillance, but it is incremental as it builds on existing deep learning methods.
The paper tackled violence detection in videos by proposing a hybrid architecture combining a 2D CNN and a BiGRU to extract spatial and temporal features, achieving up to 98% accuracy on three public datasets.
Abnormal behavior detection, action recognition, fight and violence detection in videos is an area that has attracted a lot of interest in recent years. In this work, we propose an architecture that combines a Bidirectional Gated Recurrent Unit (BiGRU) and a 2D Convolutional Neural Network (CNN) to detect violence in video sequences. A CNN is used to extract spatial characteristics from each frame, while the BiGRU extracts temporal and local motion characteristics using CNN extracted features from multiple frames. The proposed end-to-end deep learning network is tested in three public datasets with varying scene complexities. The proposed network achieves accuracies up to 98%. The obtained results are promising and show the performance of the proposed end-to-end approach.