Covariance Descriptors Meet General Vision Encoders: Riemannian Deep Learning for Medical Image Classification

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Mayr, Josef, Reithmeir, Anna, Di Folco, Maxime, Schnabel, Julia A.
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911251268894720
author Mayr, Josef
Reithmeir, Anna
Di Folco, Maxime
Schnabel, Julia A.
author_facet Mayr, Josef
Reithmeir, Anna
Di Folco, Maxime
Schnabel, Julia A.
contents Covariance descriptors capture second-order statistics of image features. They have shown strong performance in general computer vision tasks, but remain underexplored in medical imaging. We investigate their effectiveness for both conventional and learning-based medical image classification, with a particular focus on SPDNet, a classification network specifically designed for symmetric positive definite (SPD) matrices. We propose constructing covariance descriptors from features extracted by pre-trained general vision encoders (GVEs) and comparing them with handcrafted descriptors. Two GVEs - DINOv2 and MedSAM - are evaluated across eleven binary and multi-class datasets from the MedMNSIT benchmark. Our results show that covariance descriptors derived from GVE features consistently outperform those derived from handcrafted features. Moreover, SPDNet yields superior performance to state-of-the-art methods when combined with DINOv2 features. Our findings highlight the potential of combining covariance descriptors with powerful pretrained vision encoders for medical image analysis.
format Preprint
id arxiv_https___arxiv_org_abs_2511_04190
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Covariance Descriptors Meet General Vision Encoders: Riemannian Deep Learning for Medical Image Classification
Mayr, Josef
Reithmeir, Anna
Di Folco, Maxime
Schnabel, Julia A.
Computer Vision and Pattern Recognition
Covariance descriptors capture second-order statistics of image features. They have shown strong performance in general computer vision tasks, but remain underexplored in medical imaging. We investigate their effectiveness for both conventional and learning-based medical image classification, with a particular focus on SPDNet, a classification network specifically designed for symmetric positive definite (SPD) matrices. We propose constructing covariance descriptors from features extracted by pre-trained general vision encoders (GVEs) and comparing them with handcrafted descriptors. Two GVEs - DINOv2 and MedSAM - are evaluated across eleven binary and multi-class datasets from the MedMNSIT benchmark. Our results show that covariance descriptors derived from GVE features consistently outperform those derived from handcrafted features. Moreover, SPDNet yields superior performance to state-of-the-art methods when combined with DINOv2 features. Our findings highlight the potential of combining covariance descriptors with powerful pretrained vision encoders for medical image analysis.
title Covariance Descriptors Meet General Vision Encoders: Riemannian Deep Learning for Medical Image Classification
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.04190