3D Human Action Recognition with Siamese-LSTM Based Deep Metric Learning

TL;DR

This paper introduces a modular 3D human action recognition system combining Siamese-LSTM based deep metric learning with a classification module, demonstrating promising initial results and high generalizability.

Contribution

The paper presents a novel two-phase 3D action recognition system using Siamese-LSTM networks for deep metric learning, which is modular and dataset-independent.

Findings

System achieves promising initial results on standard datasets

Deep metric learning module can be trained independently of datasets

System is modular and generalizable across different datasets

Abstract

This paper proposes a new 3D Human Action Recognition system as a two-phase system: (1) Deep Metric Learning Module which learns a similarity metric between two 3D joint sequences using Siamese-LSTM networks; (2) A Multiclass Classification Module that uses the output of the first module to produce the final recognition output. This model has several advantages: the first module is trained with a larger set of data because it uses many combinations of sequence pairs.Our deep metric learning module can also be trained independently of the datasets, which makes our system modular and generalizable. We tested the proposed system on standard and newly introduced datasets that showed us that initial results are promising. We will continue developing this system by adding more sophisticated LSTM blocks and by cross-training between different datasets.