Reducing Texture Bias of Deep Neural Networks via Edge Enhancing Diffusion

arXiv:2402.09530v1 Announce Type: new
Abstract: Convolutional neural networks (CNNs) for image processing tend to focus on localized texture patterns, commonly referred to as texture bias. While most of the previous works in the literature focus on the task of image classification, we go beyond this and study the texture bias of CNNs in semantic segmentation. In this work, we propose to train CNNs on pre-processed images with less texture to reduce the texture bias. Therein, the challenge is to suppress …

abstract arxiv beyond bias classification cnns convolutional neural networks cs.cv diffusion edge focus image image processing literature networks neural networks patterns processing study texture type via