Unsupervised Object-Level Video Summarization with Online Motion Auto-Encoder
This addresses the problem of digesting and browsing videos for users by focusing on fine-grained object motions, representing an incremental advance in video summarization.
The paper tackles unsupervised object-level video summarization by extracting key motions of objects in an online manner, achieving effectiveness demonstrated on surveillance and public datasets.
Unsupervised video summarization plays an important role on digesting, browsing, and searching the ever-growing videos every day, and the underlying fine-grained semantic and motion information (i.e., objects of interest and their key motions) in online videos has been barely touched. In this paper, we investigate a pioneer research direction towards the fine-grained unsupervised object-level video summarization. It can be distinguished from existing pipelines in two aspects: extracting key motions of participated objects, and learning to summarize in an unsupervised and online manner. To achieve this goal, we propose a novel online motion Auto-Encoder (online motion-AE) framework that functions on the super-segmented object motion clips. Comprehensive experiments on a newly-collected surveillance dataset and public datasets have demonstrated the effectiveness of our proposed method.