This research paper introduces a new descriptor for human action recognition called the Distribution Action Movements Descriptor (DAM). The descriptor is based on the distribution of the movements of each joint in a skeletal representation of the user. It outperforms several state-of-the-art techniques on many well-known datasets. The descriptor is computed as a normalized histogram over a set of representative directions of the joints, obtained via clustering. The descriptor retains its temporal structure by applying a windowing scheme. This research has implications in fields such as computer vision and pattern recognition.
Publication date: 26 Oct 2023
Project Page: https://arxiv.org/abs/2310.17421v1
Paper: https://arxiv.org/pdf/2310.17421
