CVJul 13, 2022

Eliminating Gradient Conflict in Reference-based Line-Art Colorization

arXiv:2207.06095v354 citationsh-index: 22Has Code
Originality Incremental advance
AI Analysis

This work addresses a specific training challenge in computer vision for artists and designers, offering an incremental improvement over existing methods.

The paper tackles the problem of training instability in reference-based line-art colorization by identifying gradient conflicts in attention mechanisms, and proposes Stop-Gradient Attention (SGA) to alleviate this issue, resulting in significant improvements such as up to 27.21% better FID and 25.67% better SSIM on benchmarks.

Reference-based line-art colorization is a challenging task in computer vision. The color, texture, and shading are rendered based on an abstract sketch, which heavily relies on the precise long-range dependency modeling between the sketch and reference. Popular techniques to bridge the cross-modal information and model the long-range dependency employ the attention mechanism. However, in the context of reference-based line-art colorization, several techniques would intensify the existing training difficulty of attention, for instance, self-supervised training protocol and GAN-based losses. To understand the instability in training, we detect the gradient flow of attention and observe gradient conflict among attention branches. This phenomenon motivates us to alleviate the gradient issue by preserving the dominant gradient branch while removing the conflict ones. We propose a novel attention mechanism using this training strategy, Stop-Gradient Attention (SGA), outperforming the attention baseline by a large margin with better training stability. Compared with state-of-the-art modules in line-art colorization, our approach demonstrates significant improvements in Fréchet Inception Distance (FID, up to 27.21%) and structural similarity index measure (SSIM, up to 25.67%) on several benchmarks. The code of SGA is available at https://github.com/kunkun0w0/SGA .

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes