Large-scale Multiview 3D Hand Pose Dataset

TL;DR

This paper introduces a large multiview 3D hand pose dataset with detailed annotations and presents a new deep learning architecture for real-time 2D hand pose estimation, addressing data scarcity issues in the field.

Contribution

It provides a comprehensive multiview hand pose dataset with diverse annotations and proposes an effective deep learning model for real-time 2D hand pose estimation.

Findings

The dataset enables improved training of hand pose estimation models.

The proposed architecture achieves real-time performance with high accuracy.

The dataset overcomes limitations of previous datasets in size and annotation detail.

Abstract

Accurate hand pose estimation at joint level has several uses on human-robot interaction, user interfacing and virtual reality applications. Yet, it currently is not a solved problem. The novel deep learning techniques could make a great improvement on this matter but they need a huge amount of annotated data. The hand pose datasets released so far present some issues that make them impossible to use on deep learning methods such as the few number of samples, high-level abstraction annotations or samples consisting in depth maps. In this work, we introduce a multiview hand pose dataset in which we provide color images of hands and different kind of annotations for each, i.e the bounding box and the 2D and 3D location on the joints in the hand. Besides, we introduce a simple yet accurate deep learning architecture for real-time robust 2D hand pose estimation.