Multi-Action Recognition via Stochastic Modelling of Optical Flow and Gradients

TL;DR

This paper introduces a simpler joint segmentation and classification method for multi-action recognition using Gaussian mixture models on low-dimensional features, achieving higher accuracy than previous HMM-based methods.

Contribution

The paper presents a novel, less complex approach that combines segmentation and classification with Gaussian mixture models for multi-action recognition.

Findings

Achieved 78.3% accuracy on KTH dataset

Outperformed recent HMM-based approaches

Simplified the recognition process

Abstract

In this paper we propose a novel approach to multi-action recognition that performs joint segmentation and classification. This approach models each action using a Gaussian mixture using robust low-dimensional action features. Segmentation is achieved by performing classification on overlapping temporal windows, which are then merged to produce the final result. This approach is considerably less complicated than previous methods which use dynamic programming or computationally expensive hidden Markov models (HMMs). Initial experiments on a stitched version of the KTH dataset show that the proposed approach achieves an accuracy of 78.3%, outperforming a recent HMM-based approach which obtained 71.2%.