Hand Image Understanding via Deep Multi-Task Learning

TL;DR

This paper introduces a multi-task learning framework that jointly estimates hand pose, segmentation, and 3D information from a single RGB image, significantly improving accuracy in challenging scenarios.

Contribution

The novel HIU framework integrates multiple hand understanding tasks with a cascaded multi-task learning approach and self-supervised strategies, advancing single-image hand analysis.

Findings

Outperforms state-of-the-art methods on multiple datasets.

Capable of recovering detailed hand mesh in difficult conditions.

Effectively integrates multiple hand-related tasks in a unified model.

Abstract

Analyzing and understanding hand information from multimedia materials like images or videos is important for many real world applications and remains active in research community. There are various works focusing on recovering hand information from single image, however, they usually solve a single task, for example, hand mask segmentation, 2D/3D hand pose estimation, or hand mesh reconstruction and perform not well in challenging scenarios. To further improve the performance of these tasks, we propose a novel Hand Image Understanding (HIU) framework to extract comprehensive information of the hand object from a single RGB image, by jointly considering the relationships between these tasks. To achieve this goal, a cascaded multi-task learning (MTL) backbone is designed to estimate the 2D heat maps, to learn the segmentation mask, and to generate the intermediate 3D information encoding, followed by a coarse-to-fine learning paradigm and a self-supervised learning strategy. Qualitative experiments demonstrate that our approach is capable of recovering reasonable mesh representations even in challenging situations. Quantitatively, our method significantly outperforms the state-of-the-art approaches on various widely-used datasets, in terms of diverse evaluation metrics.