Learning Pixel-adaptive Multi-layer Perceptrons for Real-time Image Enhancement

TL;DR

This paper introduces a novel pixel-adaptive MLP framework that combines bilateral grids with non-linear mappings for real-time, high-quality image enhancement, overcoming limitations of previous linear and global methods.

Contribution

The proposed BPAM framework uniquely integrates bilateral grids with MLPs, enabling localized, non-linear color transformations for image enhancement in real-time.

Findings

Outperforms state-of-the-art methods in image enhancement quality

Maintains real-time processing speeds

Effectively models complex color relationships

Abstract

Deep learning-based bilateral grid processing has emerged as a promising solution for image enhancement, inherently encoding spatial and intensity information while enabling efficient full-resolution processing through slicing operations. However, existing approaches are limited to linear affine transformations, hindering their ability to model complex color relationships. Meanwhile, while multi-layer perceptrons (MLPs) excel at non-linear mappings, traditional MLP-based methods employ globally shared parameters, which is hard to deal with localized variations. To overcome these dual challenges, we propose a Bilateral Grid-based Pixel-Adaptive Multi-layer Perceptron (BPAM) framework. Our approach synergizes the spatial modeling of bilateral grids with the non-linear capabilities of MLPs. Specifically, we generate bilateral grids containing MLP parameters, where each pixel dynamically retrieves its unique transformation parameters and obtain a distinct MLP for color mapping based on spatial coordinates and intensity values. In addition, we propose a novel grid decomposition strategy that categorizes MLP parameters into distinct types stored in separate subgrids. Multi-channel guidance maps are used to extract category-specific parameters from corresponding subgrids, ensuring effective utilization of color information during slicing while guiding precise parameter generation. Extensive experiments on public datasets demonstrate that our method outperforms state-of-the-art methods in performance while maintaining real-time processing capabilities.