Balancing Accuracy and Training Time in Federated Learning for Violence Detection in Surveillance Videos: A Study of Neural Network Architectures

TL;DR

This study explores neural network architectures for violence detection in surveillance videos within federated learning, proposing modifications and techniques to balance accuracy and training time, achieving improved results over existing models.

Contribution

Introduces the Diff-Gated architecture and a federated adaptation method, enhancing violence detection accuracy while managing training efficiency in federated settings.

Findings

Diff-Gated outperforms baseline models in accuracy.

Federated training achieves comparable results to centralized models.

Super-convergence and transfer learning improve training efficiency.

Abstract

This paper presents an investigation into machine learning techniques for violence detection in videos and their adaptation to a federated learning context. The study includes experiments with spatio-temporal features extracted from benchmark video datasets, comparison of different methods, and proposal of a modified version of the "Flow-Gated" architecture called "Diff-Gated." Additionally, various machine learning techniques, including super-convergence and transfer learning, are explored, and a method for adapting centralized datasets to a federated learning context is developed. The research achieves better accuracy results compared to state-of-the-art models by training the best violence detection model in a federated learning context.