Quantifying and Inducing Shape Bias in CNNs via Max-Pool Dilation

• URL: http://arxiv.org/abs/2601.05599v1
• Date: Fri, 09 Jan 2026 07:36:29 GMT
• Title: Quantifying and Inducing Shape Bias in CNNs via Max-Pool Dilation
• Authors: Takito Sawada, Akinori Iwata, Masahiro Okuda,
• Abstract summary: Convolutional Neural Networks (CNNs) are known to exhibit a strong texture bias, favoring local patterns over global shape information.<n>We propose a data-driven metric that quantifies the shape-texture balance of a dataset by computing the Structural Similarity Index (SSIM)<n>We introduce a computationally efficient adaptation method that promotes shape bias by modifying the dilation of max-pooling operations while keeping convolutional weights frozen.
• Score: 0.9176056742068813
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Convolutional Neural Networks (CNNs) are known to exhibit a strong texture bias, favoring local patterns over global shape information--a tendency inherent to their convolutional architecture. While this bias is beneficial for texture-rich natural images, it often degrades performance on shape-dominant data such as illustrations and sketches. Although prior work has proposed shape-biased models to mitigate this issue, these approaches lack a quantitative metric for identifying which datasets would actually benefit from such modifications. To address this gap, we propose a data-driven metric that quantifies the shape-texture balance of a dataset by computing the Structural Similarity Index (SSIM) between each image's luminance channel and its L0-smoothed counterpart. Building on this metric, we further introduce a computationally efficient adaptation method that promotes shape bias by modifying the dilation of max-pooling operations while keeping convolutional weights frozen. Experimental results show that this approach consistently improves classification accuracy on shape-dominant datasets, particularly in low-data regimes where full fine-tuning is impractical, requiring training only the final classification layer.

Related papers

• Nonparametric Data Attribution for Diffusion Models [57.820618036556084]
  Data attribution for generative models seeks to quantify the influence of individual training examples on model outputs.<n>We propose a nonparametric attribution method that operates entirely on data, measuring influence via patch-level similarity between generated and training images.
  arXiv  Detail & Related papers  (2025-10-16T03:37:16Z)
• Fusing CFD and measurement data using transfer learning [49.1574468325115]
  We introduce a non-linear method based on neural networks combining simulation and measurement data via transfer learning.<n>In a first step, the neural network is trained on simulation data to learn spatial features of the distributed quantities.<n>The second step involves transfer learning on the measurement data to correct for systematic errors between simulation and measurement by only re-training a small subset of the entire neural network model.
  arXiv  Detail & Related papers  (2025-07-28T07:21:46Z)
• GratNet: A Photorealistic Neural Shader for Diffractive Surfaces [0.0]
  We present a multi-layer perceptron (MLP) based method for data-driven rendering of diffractive surfaces with high accuracy and efficiency.<n>We demonstrate the high-quality reconstruction of the ground-truth using Peak-Signal-to-Noise (PSNR), Structural Similarity Index Measure (SSIM) and a flipping difference evaluator (FLIP) as evaluation metrics.
  arXiv  Detail & Related papers  (2025-06-18T18:58:00Z)
• Mean flow data assimilation using physics-constrained Graph Neural Networks [0.0]
  This study introduces a novel data assimilation approach that integrates Graph Neural Networks (GNNs) with optimisation techniques to enhance the accuracy of mean flow reconstruction.<n>The GNN framework is well-suited for handling unstructured data, which is common in the complex geometries encountered in Computational Fluid Dynamics (CFD)<n>Results demonstrate significant improvements in the accuracy of mean flow reconstructions, even with limited training data, compared to analogous purely data-driven models.
  arXiv  Detail & Related papers  (2024-11-14T14:31:52Z)
• Distance Weighted Trans Network for Image Completion [52.318730994423106]
  We propose a new architecture that relies on Distance-based Weighted Transformer (DWT) to better understand the relationships between an image's components. CNNs are used to augment the local texture information of coarse priors. DWT blocks are used to recover certain coarse textures and coherent visual structures.
  arXiv  Detail & Related papers  (2023-10-11T12:46:11Z)
• Topology Estimation of Simulated 4D Image Data by Combining Downscaling and Convolutional Neural Networks [0.0]
  This study estimates the type of 4D image-type data cubes that exhibit topological intricateness and size above our current processing capacity.<n>Experiments using synthesised 4D data and a real-world 3D data set demonstrate that it is possible to circumvent computational complexity issues.
  arXiv  Detail & Related papers  (2023-06-26T06:13:43Z)
• Minimizing the Accumulated Trajectory Error to Improve Dataset Distillation [151.70234052015948]
  We propose a novel approach that encourages the optimization algorithm to seek a flat trajectory. We show that the weights trained on synthetic data are robust against the accumulated errors perturbations with the regularization towards the flat trajectory. Our method, called Flat Trajectory Distillation (FTD), is shown to boost the performance of gradient-matching methods by up to 4.7%.
  arXiv  Detail & Related papers  (2022-11-20T15:49:11Z)
• Neural Jacobian Fields: Learning Intrinsic Mappings of Arbitrary Meshes [38.157373733083894]
  This paper introduces a framework designed to accurately predict piecewise linear mappings of arbitrary meshes via a neural network. The framework is based on reducing the neural aspect to a prediction of a matrix for a single point, conditioned on a global shape descriptor. By operating in the intrinsic gradient domain of each individual mesh, it allows the framework to predict highly-accurate mappings.
  arXiv  Detail & Related papers  (2022-05-05T19:51:13Z)
• Bending Graphs: Hierarchical Shape Matching using Gated Optimal Transport [80.64516377977183]
  Shape matching has been a long-studied problem for the computer graphics and vision community. We investigate a hierarchical learning design, to which we incorporate local patch-level information and global shape-level structures. We propose a novel optimal transport solver by recurrently updating features on non-confident nodes to learn globally consistent correspondences between the shapes.
  arXiv  Detail & Related papers  (2022-02-03T11:41:46Z)
• Out-of-distribution Generalization via Partial Feature Decorrelation [72.96261704851683]
  We present a novel Partial Feature Decorrelation Learning (PFDL) algorithm, which jointly optimize a feature decomposition network and the target image classification model. The experiments on real-world datasets demonstrate that our method can improve the backbone model's accuracy on OOD image classification datasets.
  arXiv  Detail & Related papers  (2020-07-30T05:48:48Z)

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.