LP-IOANet: Efficient High Resolution Document Shadow Removal
Konstantinos Georgiadis, Mehmet Kerim Yucel, Evangelos Skartados, Valia Dimaridou, Anastasios Drosou, Albert Saà-Garriga, Bruno Manganelli
Abstract
Document shadow removal is an integral task in document enhancement pipelines, as it improves visibility, readability and thus the overall quality. Assuming that the majority of practical document shadow removal scenarios require real-time, accurate models that can produce high-resolution outputs in-the-wild, we propose Laplacian Pyramid with Input/Output Attention Network (LP-IOANet), a novel pipeline with a lightweight architecture and an upsampling module. Furthermore, we propose three new datasets which cover a wide range of lighting conditions, images, shadow shapes and viewpoints. Our results show that we outperform the state-of-the-art by a 35% relative improvement in mean average error (MAE), while running real-time in four times the resolution (of the state-of-the-art method) on a mobile device.
BibTeX
@inproceedings{icassp2023_lpioanetefficien,
title = {LP-IOANet: Efficient High Resolution Document Shadow Removal},
author = {Konstantinos Georgiadis and Mehmet Kerim Yucel and Evangelos Skartados and Valia Dimaridou and Anastasios Drosou and Albert Saà-Garriga and Bruno Manganelli},
booktitle = {ICASSP 2023},
year = {2023}
}