Shape-Based Inductive Bias for Glioma Grading from Tumor Contours

2026/07/27 by Puneet Velidi, Michelle F. Miranda, Farouk Nathoo +2
Computer Science · Engineering · #cs.CV #cs.LG #eess.IV

paper · pdf

arxiv created 2026/07/27 · arxiv updated 2026/07/30

Abstract

Glioma grading from tumor contours is often treated as a pixel problem even when the signal of interest is shape. We align closed contours with a functional shape-alignment framework, separate global deformation from residual Fourier shape, and organize these quantities as frequency-ordered tokens. In five-fold patient-disjoint cross-validation on BraTS~2020 tumor contours, with model selection performed using grouped inner validation, a compact multilayer perceptron (MLP) achieves the highest mean balanced accuracy at 71.5%, compared with 65.9% for ResNet-18 and 63.3% for ViT-Tiny. It also gives the highest mean low-grade glioma F1 at 54.9%. Its pooled out-of-fold balanced accuracy is 72.4% (patient-bootstrap 95% CI: 66.4--77.8%). The selected MLPs use 2.9k--117.3k parameters across folds, at least 46 times fewer than the pixel baselines. In a controlled noise-free simulation, shape-based models reach 56.3--71.5% balanced accuracy while the pixel models remain at 50.0--52.5%. This work demonstrates how incorporating a shape-based inductive bias at the representation level can improve interpretability and scalability while enabling substantial dimensionality reduction.

Citations

• Multivariate Planar Curves: A Statistical Framework for Shape Analysis in Images
• A Functional Approach to Curve Alignment and Shape Analysis
• Functional Autoencoder for Smoothing and Representation Learning
• Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges
• UNETR: Transformers for 3D Medical Image Segmentation
• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
• ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness
• Attention Is All You Need
• Deep Residual Learning for Image Recognition
• Radiomics: Images Are More than Pictures, They Are Data
• U-Net: Convolutional Networks for Biomedical Image Segmentation
• The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS)
• Radiomics: Extracting more information from medical images using advanced feature analysis
• Fourier Descriptors for Plane Closed Curves
• BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

Related

• CONSeg: Voxelwise Glioma Conformal Segmentation
• Uncertainty-driven refinement of tumor-core segmentation using 3D-to-2D networks with label uncertainty
• Rethinking Radiology: An Analysis of Different Approaches to BraTS
• Spherical coordinates transformation pre-processing in Deep Convolution Neural Networks for brain tumor segmentation in MRI
• WHO 2016 subtyping and automated segmentation of glioma using multi-task deep learning