# DeepHuMS: Deep Human Motion Signature for 3D Skeletal Sequences

## TL;DR

DeepHuMS introduces a deep learning-based 3D human motion descriptor that improves retrieval accuracy and generalization across large datasets, addressing challenges like noise, missing data, and varying motion speeds.

## Contribution

It proposes a novel learned embedding for 3D human motion that outperforms traditional methods and enables sub-motion search, validated on large-scale datasets.

## Key findings

- Superior performance on NTU RGB+D and HDM05 datasets
- Effective in noisy and real-world data conditions
- Enables sub-motion retrieval using learned embeddings

## Abstract

3D Human Motion Indexing and Retrieval is an interesting problem due to the rise of several data-driven applications aimed at analyzing and/or re-utilizing 3D human skeletal data, such as data-driven animation, analysis of sports bio-mechanics, human surveillance etc. Spatio-temporal articulations of humans, noisy/missing data, different speeds of the same motion etc. make it challenging and several of the existing state of the art methods use hand-craft features along with optimization based or histogram based comparison in order to perform retrieval. Further, they demonstrate it only for very small datasets and few classes. We make a case for using a learned representation that should recognize the motion as well as enforce a discriminative ranking. To that end, we propose, a 3D human motion descriptor learned using a deep network. Our learned embedding is generalizable and applicable to real-world data - addressing the aforementioned challenges and further enables sub-motion searching in its embedding space using another network. Our model exploits the inter-class similarity using trajectory cues, and performs far superior in a self-supervised setting. State of the art results on all these fronts is shown on two large scale 3D human motion datasets - NTU RGB+D and HDM05.