Single Run Action Detector over Video Stream -- A Privacy Preserving Approach
This work addresses the problem of privacy-preserving activity monitoring for users concerned about data security, offering a solution that can run efficiently on edge devices.
This paper introduces Single Run Action Detector (S-RAD), a real-time privacy-preserving action detector that performs end-to-end action localization and classification. It achieves comparable accuracy to State-of-the-Art approaches on UCF-Sports and UR Fall datasets, while significantly reducing model size and computational demand, enabling real-time execution on edge devices.
This paper takes initial strides at designing and evaluating a vision-based system for privacy ensured activity monitoring. The proposed technology utilizing Artificial Intelligence (AI)-empowered proactive systems offering continuous monitoring, behavioral analysis, and modeling of human activities. To this end, this paper presents Single Run Action Detector (S-RAD) which is a real-time privacy-preserving action detector that performs end-to-end action localization and classification. It is based on Faster-RCNN combined with temporal shift modeling and segment based sampling to capture the human actions. Results on UCF-Sports and UR Fall dataset present comparable accuracy to State-of-the-Art approaches with significantly lower model size and computation demand and the ability for real-time execution on edge embedded device (e.g. Nvidia Jetson Xavier).