Normalize Filters! Classical Wisdom for Deep Vision

• URL: http://arxiv.org/abs/2506.04401v1
• Date: Wed, 04 Jun 2025 19:32:42 GMT
• Title: Normalize Filters! Classical Wisdom for Deep Vision
• Authors: Gustavo Perez, Stella X. Yu,
• Abstract summary: We propose filter normalization, followed by learnable scaling and shifting, akin to batch normalization.<n>Our method achieves significant improvements on artificial and natural intensity variation benchmarks.<n>Our analysis reveals that unnormalized filters degrade performance, whereas filter normalization regularizes learning.
• Score: 32.953265823087754
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Classical image filters, such as those for averaging or differencing, are carefully normalized to ensure consistency, interpretability, and to avoid artifacts like intensity shifts, halos, or ringing. In contrast, convolutional filters learned end-to-end in deep networks lack such constraints. Although they may resemble wavelets and blob/edge detectors, they are not normalized in the same or any way. Consequently, when images undergo atmospheric transfer, their responses become distorted, leading to incorrect outcomes. We address this limitation by proposing filter normalization, followed by learnable scaling and shifting, akin to batch normalization. This simple yet effective modification ensures that the filters are atmosphere-equivariant, enabling co-domain symmetry. By integrating classical filtering principles into deep learning (applicable to both convolutional neural networks and convolution-dependent vision transformers), our method achieves significant improvements on artificial and natural intensity variation benchmarks. Our ResNet34 could even outperform CLIP by a large margin. Our analysis reveals that unnormalized filters degrade performance, whereas filter normalization regularizes learning, promotes diversity, and improves robustness and generalization.

Related papers

• Catch-Up Mix: Catch-Up Class for Struggling Filters in CNN [15.3232203753165]
  Deep learning models often face challenges related to complexity and overfitting. One notable concern is that the model often relies heavily on a limited subset of filters for making predictions. We present a novel method called Catch-up Mix, which provides learning opportunities to a wide range of filters during training.
  arXiv  Detail & Related papers  (2024-01-24T02:42:50Z)
• Understanding the Covariance Structure of Convolutional Filters [86.0964031294896]
  Recent ViT-inspired convolutional networks such as ConvMixer and ConvNeXt use large-kernel depthwise convolutions with notable structure. We first observe that such learned filters have highly-structured covariance matrices, and we find that covariances calculated from small networks may be used to effectively initialize a variety of larger networks.
  arXiv  Detail & Related papers  (2022-10-07T15:59:13Z)
• Combinations of Adaptive Filters [38.0505909175152]
  Combination of adaptive filters exploits divide and conquer principle. In particular, the problem of combining the outputs of several learning algorithms has been studied in the computational learning field.
  arXiv  Detail & Related papers  (2021-12-22T22:21:43Z)
• Learning Versatile Convolution Filters for Efficient Visual Recognition [125.34595948003745]
  This paper introduces versatile filters to construct efficient convolutional neural networks. We conduct theoretical analysis on network complexity and an efficient convolution scheme is introduced. Experimental results on benchmark datasets and neural networks demonstrate that our versatile filters are able to achieve comparable accuracy as that of original filters.
  arXiv  Detail & Related papers  (2021-09-20T06:07:14Z)
• Unsharp Mask Guided Filtering [53.14430987860308]
  The goal of this paper is guided image filtering, which emphasizes the importance of structure transfer during filtering. We propose a new and simplified formulation of the guided filter inspired by unsharp masking. Our formulation enjoys a filtering prior to a low-pass filter and enables explicit structure transfer by estimating a single coefficient.
  arXiv  Detail & Related papers  (2021-06-02T19:15:34Z)
• Efficient Spatially Adaptive Convolution and Correlation [11.167305713900074]
  We provide a representation-theoretic framework that allows for spatially varying linear transformations to be applied to the filter. This framework allows for efficient implementation of extended convolution and correlation for transformation groups such as rotation (in 2D and 3D) and scale. We present applications to pattern matching, image feature description, vector field visualization, and adaptive image filtering.
  arXiv  Detail & Related papers  (2020-06-23T17:41:10Z)
• Dependency Aware Filter Pruning [74.69495455411987]
  Pruning a proportion of unimportant filters is an efficient way to mitigate the inference cost. Previous work prunes filters according to their weight norms or the corresponding batch-norm scaling factors. We propose a novel mechanism to dynamically control the sparsity-inducing regularization so as to achieve the desired sparsity.
  arXiv  Detail & Related papers  (2020-05-06T07:41:22Z)
• Filter Grafting for Deep Neural Networks: Reason, Method, and Cultivation [86.91324735966766]
  Filter is the key component in modern convolutional neural networks (CNNs) In this paper, we introduce filter grafting (textbfMethod) to achieve this goal. We develop a novel criterion to measure the information of filters and an adaptive weighting strategy to balance the grafted information among networks.
  arXiv  Detail & Related papers  (2020-04-26T08:36:26Z)
• Recognizing Instagram Filtered Images with Feature De-stylization [81.38905784617089]
  This paper presents a study on how popular pretrained models are affected by commonly used Instagram filters. Our analysis suggests that simple structure preserving filters which only alter the global appearance of an image can lead to large differences in the convolutional feature space. We introduce a lightweight de-stylization module that predicts parameters used for scaling and shifting feature maps to "undo" the changes incurred by filters.
  arXiv  Detail & Related papers  (2019-12-30T16:48:16Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.