Towards Hard and Soft Shadow Removal via Dual-Branch Separation Network and Vision Transformer

TL;DR

This paper introduces a dual-branch network combining Vision Transformer and UNet++ to specifically address hard and soft shadow removal, significantly improving accuracy over existing methods.

Contribution

It presents a novel dual-path model that classifies and separately processes hard and soft shadows using specialized loss functions and a hybrid transformer-based architecture.

Findings

Achieves 2.905 RMSE on ISTD dataset, outperforming state-of-the-art methods.

Effectively classifies and removes both hard and soft shadows.

Enhances edge detail and feature fusion in shadow removal.

Abstract

Image shadow removal is a crucial task in computer vision. In real-world scenes, shadows alter image color and brightness, posing challenges for perception and texture recognition. Traditional and deep learning methods often overlook the distinct needs for handling hard and soft shadows, thereby lacking detailed processing to specifically address each type of shadow in images.We propose a dual-path model that processes these shadows separately using specially designed loss functions to accomplish the hard and soft shadow removal. The model classifies shadow types and processes them through appropriate paths to produce shadow-free outputs, integrating a Vision Transformer with UNet++ for enhanced edge detail and feature fusion. Our model outperforms state-of-the-art methods and achieves 2.905 RMSE value on the ISTD dataset, which demonstrates greater effectiveness than typical single-path approaches.