A Benchmark for Inpainting of Clothing Images with Irregular Holes

TL;DR

This paper introduces a comprehensive benchmark for clothing image inpainting, proposing dilated partial convolutions that improve performance, especially for larger masks, advancing fashion image understanding.

Contribution

It presents a new benchmark dataset for clothing image inpainting and introduces dilated partial convolutions to enhance inpainting efficiency and accuracy.

Findings

Dilated partial convolutions outperform other inpainting methods.

Performance improves with larger masks, especially over 20% of the image.

The proposed method reduces the number of layers needed for effective inpainting.

Abstract

Fashion image understanding is an active research field with a large number of practical applications for the industry. Despite its practical impacts on intelligent fashion analysis systems, clothing image inpainting has not been extensively examined yet. For that matter, we present an extensive benchmark of clothing image inpainting on well-known fashion datasets. Furthermore, we introduce the use of a dilated version of partial convolutions, which efficiently derive the mask update step, and empirically show that the proposed method reduces the required number of layers to form fully-transparent masks. Experiments show that dilated partial convolutions (DPConv) improve the quantitative inpainting performance when compared to the other inpainting strategies, especially it performs better when the mask size is 20% or more of the image. \keywords{image inpainting, fashion image understanding, dilated convolutions, partial convolutions