Learning Higher-order Object Interactions for Keypoint-based Video Understanding
This addresses the problem of efficient action recognition for AR/VR applications by reducing computation while maintaining accuracy, though it is incremental in using object keypoints to enhance existing keypoint-based methods.
The paper tackles action recognition in video by using only keypoint data to model higher-order interactions between human and object keypoints, achieving tracking and classification at 5 FPS and recovering context loss on AVA and Kinetics datasets.
Action recognition is an important problem that requires identifying actions in video by learning complex interactions across scene actors and objects. However, modern deep-learning based networks often require significant computation, and may capture scene context using various modalities that further increases compute costs. Efficient methods such as those used for AR/VR often only use human-keypoint information but suffer from a loss of scene context that hurts accuracy. In this paper, we describe an action-localization method, KeyNet, that uses only the keypoint data for tracking and action recognition. Specifically, KeyNet introduces the use of object based keypoint information to capture context in the scene. Our method illustrates how to build a structured intermediate representation that allows modeling higher-order interactions in the scene from object and human keypoints without using any RGB information. We find that KeyNet is able to track and classify human actions at just 5 FPS. More importantly, we demonstrate that object keypoints can be modeled to recover any loss in context from using keypoint information over AVA action and Kinetics datasets.