Enhancing convolutional neural network generalizability via low-rank weight approximation

TL;DR

This paper introduces a self-supervised image denoising framework using Tucker low-rank tensor approximation, which requires fewer parameters and training data, leading to improved generalizability and competitive performance.

Contribution

The paper proposes a novel self-supervised denoising method based on Tucker low-rank tensor approximation that trains on a single image, reducing data needs and enhancing model generalizability.

Findings

Outperforms existing non-learning-based denoising methods.

Achieves comparable results to supervised methods like DnCNN.

Effective on both synthetic and real-world noisy images.

Abstract

Noise is ubiquitous during image acquisition. Sufficient denoising is often an important first step for image processing. In recent decades, deep neural networks (DNNs) have been widely used for image denoising. Most DNN-based image denoising methods require a large-scale dataset or focus on supervised settings, in which single/pairs of clean images or a set of noisy images are required. This poses a significant burden on the image acquisition process. Moreover, denoisers trained on datasets of limited scale may incur over-fitting. To mitigate these issues, we introduce a new self-supervised framework for image denoising based on the Tucker low-rank tensor approximation. With the proposed design, we are able to characterize our denoiser with fewer parameters and train it based on a single image, which considerably improves the model's generalizability and reduces the cost of data acquisition. Extensive experiments on both synthetic and real-world noisy images have been conducted. Empirical results show that our proposed method outperforms existing non-learning-based methods (e.g., low-pass filter, non-local mean), single-image unsupervised denoisers (e.g., DIP, NN+BM3D) evaluated on both in-sample and out-sample datasets. The proposed method even achieves comparable performances with some supervised methods (e.g., DnCNN).