I2UV-HandNet: Image-to-UV Prediction Network for Accurate and High-fidelity 3D Hand Mesh Modeling

TL;DR

This paper introduces I2UV-HandNet, a novel UV-based 3D hand shape representation and a deep learning framework that significantly improves the accuracy and fidelity of 3D hand reconstruction from RGB images, even under occlusions.

Contribution

It presents the first UV-based 3D hand shape representation and a two-stage network for high-fidelity hand mesh reconstruction from images.

Findings

Achieves state-of-the-art results on multiple benchmarks.

Demonstrates the effectiveness of UV-based representation for hand modeling.

Provides a new approach for high-resolution 3D hand reconstruction.

Abstract

Reconstructing a high-precision and high-fidelity 3D human hand from a color image plays a central role in replicating a realistic virtual hand in human-computer interaction and virtual reality applications. The results of current methods are lacking in accuracy and fidelity due to various hand poses and severe occlusions. In this study, we propose an I2UV-HandNet model for accurate hand pose and shape estimation as well as 3D hand super-resolution reconstruction. Specifically, we present the first UV-based 3D hand shape representation. To recover a 3D hand mesh from an RGB image, we design an AffineNet to predict a UV position map from the input in an image-to-image translation fashion. To obtain a higher fidelity shape, we exploit an additional SRNet to transform the low-resolution UV map outputted by AffineNet into a high-resolution one. For the first time, we demonstrate the characterization capability of the UV-based hand shape representation. Our experiments show that the proposed method achieves state-of-the-art performance on several challenging benchmarks.