A discussion on the validation tests employed to compare human action recognition methods using the MSR Action3D dataset

TL;DR

This paper reviews human action recognition methods using the MSR Action3D dataset, highlighting the inconsistency in validation methods and proposing rankings based on different validation approaches to clarify comparisons.

Contribution

It provides a comprehensive review of existing validation methods in MSR Action3D-based recognition studies and offers a structured comparison to address inconsistencies.

Findings

Validation methods vary widely across studies

Most works compare results without considering validation differences

Proposed rankings clarify the impact of validation approaches

Abstract

This paper aims to determine which is the best human action recognition method based on features extracted from RGB-D devices, such as the Microsoft Kinect. A review of all the papers that make reference to MSR Action3D, the most used dataset that includes depth information acquired from a RGB-D device, has been performed. We found that the validation method used by each work differs from the others. So, a direct comparison among works cannot be made. However, almost all the works present their results comparing them without taking into account this issue. Therefore, we present different rankings according to the methodology used for the validation in orden to clarify the existing confusion.