Online Detection of Action Start in Untrimmed, Streaming Videos
We aim to tackle a novel task in action detection - Online Detection of Action Start (ODAS) in untrimmed, streaming videos. The goal of ODAS is to detect the start of an action instance, with high cat- egorization accuracy and low detection latency. ODAS is important in many applications such as early alert generation to allow timely secu- rity or emergency response. We propose three novel methods to specifi- cally address the challenges in training ODAS models: (1) hard negative samples generation based on Generative Adversarial Network (GAN) to distinguish ambiguous background, (2) explicitly modeling the tempo- ral consistency between data around action start and data succeeding action start, and (3) adaptive sampling strategy to handle the scarcity of training data. We conduct extensive experiments using THUMOS’14 and ActivityNet. We show that our proposed methods lead to signifi- cant performance gains and improve the state-of-the-art methods. An ablation study confirms the effectiveness of each proposed method.