Wavelet Transform-assisted Adaptive Generative Modeling for Colorization

TL;DR

This paper introduces a wavelet transform-assisted generative model for image colorization that leverages multi-scale wavelet features and dual consistency constraints to improve colorization quality, robustness, and diversity.

Contribution

It proposes a novel wavelet domain score-based generative model with dual consistency terms, enhancing unsupervised image colorization performance.

Findings

Significant improvement in colorization quality and diversity.

Enhanced robustness against colorization ambiguities.

Effective reduction of data manifold dimension.

Abstract

Unsupervised deep learning has recently demonstrated the promise of producing high-quality samples. While it has tremendous potential to promote the image colorization task, the performance is limited owing to the high-dimension of data manifold and model capability. This study presents a novel scheme that exploits the score-based generative model in wavelet domain to address the issues. By taking advantage of the multi-scale and multi-channel representation via wavelet transform, the proposed model learns the richer priors from stacked coarse and detailed wavelet coefficient components jointly and effectively. This strategy also reduces the dimension of the original manifold and alleviates the curse of dimensionality, which is beneficial for estimation and sampling. Moreover, dual consistency terms in the wavelet domain, namely data-consistency and structure-consistency are devised to leverage colorization task better. Specifically, in the training phase, a set of multi-channel tensors consisting of wavelet coefficients is used as the input to train the network with denoising score matching. In the inference phase, samples are iteratively generated via annealed Langevin dynamics with data and structure consistencies. Experiments demonstrated remarkable improvements of the proposed method on both generation and colorization quality, particularly in colorization robustness and diversity.