Do Segmentation Models Understand Vascular Structure? A Blob-Based XAI Framework

• URL: http://arxiv.org/abs/2504.11469v1
• Date: Fri, 11 Apr 2025 12:42:52 GMT
• Title: Do Segmentation Models Understand Vascular Structure? A Blob-Based XAI Framework
• Authors: Guillaume Garret, Antoine Vacavant, Carole Frindel,
• Abstract summary: Deep learning models have achieved impressive performance in medical image segmentation, yet their black-box nature limits clinical adoption.<n>We present a novel explainability pipeline for 3D vessel segmentation, combining gradient-based attribution with graph-guided point selection.<n>Our results underline the importance of structured explainability tools and highlight the current limitations of segmentation models in capturing global vascular context.
• Score: 0.3277163122167433
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Deep learning models have achieved impressive performance in medical image segmentation, yet their black-box nature limits clinical adoption. In vascular applications, trustworthy segmentation should rely on both local image cues and global anatomical structures, such as vessel connectivity or branching. However, the extent to which models leverage such global context remains unclear. We present a novel explainability pipeline for 3D vessel segmentation, combining gradient-based attribution with graph-guided point selection and a blob-based analysis of Saliency maps. Using vascular graphs extracted from ground truth, we define anatomically meaningful points of interest (POIs) and assess the contribution of input voxels via Saliency maps. These are analyzed at both global and local scales using a custom blob detector. Applied to IRCAD and Bullitt datasets, our analysis shows that model decisions are dominated by highly localized attribution blobs centered near POIs. Attribution features show little correlation with vessel-level properties such as thickness, tubularity, or connectivity -- suggesting limited use of global anatomical reasoning. Our results underline the importance of structured explainability tools and highlight the current limitations of segmentation models in capturing global vascular context.

Related papers

• PathSegDiff: Pathology Segmentation using Diffusion model representations [63.20694440934692]
  We propose PathSegDiff, a novel approach for histopathology image segmentation that leverages Latent Diffusion Models (LDMs) as pre-trained featured extractors.<n>Our method utilizes a pathology-specific LDM, guided by a self-supervised encoder, to extract rich semantic information from H&E stained histopathology images.<n>Our experiments demonstrate significant improvements over traditional methods on the BCSS and GlaS datasets.
  arXiv  Detail & Related papers  (2025-04-09T14:58:21Z)
• From Pixels to Histopathology: A Graph-Based Framework for Interpretable Whole Slide Image Analysis [81.19923502845441]
  We develop a graph-based framework that constructs WSI graph representations.<n>We build tissue representations (nodes) that follow biological boundaries rather than arbitrary patches.<n>In our method's final step, we solve the diagnostic task through a graph attention network.
  arXiv  Detail & Related papers  (2025-03-14T20:15:04Z)
• Optimized Vessel Segmentation: A Structure-Agnostic Approach with Small Vessel Enhancement and Morphological Correction [7.882674026364302]
  We propose a structure-agnostic approach incorporating small vessel enhancement and morphological correction for multi-modality vessel segmentation. Our approach achieves superior segmentation accuracy, generalization, and a 34.6% improvement in connectivity, underscoring its clinical potential.
  arXiv  Detail & Related papers  (2024-11-22T08:38:30Z)
• KLDD: Kalman Filter based Linear Deformable Diffusion Model in Retinal Image Segmentation [51.03868117057726]
  This paper proposes a novel Kalman filter based Linear Deformable Diffusion (KLDD) model for retinal vessel segmentation. Our model employs a diffusion process that iteratively refines the segmentation, leveraging the flexible receptive fields of deformable convolutions. Experiments are evaluated on retinal fundus image datasets (DRIVE, CHASE_DB1) and the 3mm and 6mm of the OCTA-500 dataset.
  arXiv  Detail & Related papers  (2024-09-19T14:21:38Z)
• Fitting tree model with CNN and geodesics to track vesselsand application to Ultrasound Localization Microscopy data [5.1082516810678396]
  We build a model to carry tracking on Ultrasound Localization Microscopy (ULM) data. We also test our framework on synthetic and eye fundus data. Results show that scarcity of well annotated ULM data is an obstacle to localization of vascular landmarks.
  arXiv  Detail & Related papers  (2023-11-13T09:25:03Z)
• GAEI-UNet: Global Attention and Elastic Interaction U-Net for Vessel Image Segmentation [0.0]
  Vessel image segmentation plays a pivotal role in medical diagnostics, aiding in the early detection and treatment of vascular diseases. We propose GAEI-UNet, a novel model that combines global attention and elastic interaction-based techniques. By capturing the forces generated by misalignment between target and predicted shapes, our model effectively learns to preserve the correct topology of vessel networks.
  arXiv  Detail & Related papers  (2023-08-16T13:10:32Z)
• Morphology Edge Attention Network and Optimal Geometric Matching Connection model for vascular segmentation [3.6368619769561668]
  We propose a novel Morphology Edge Attention Network (MEA-Net) for the segmentation of vessel-like structures. We also present an Optimal Geometric Matching Connection (OGMC) model to connect the broken vessel segments. Our method achieves superior or competitive results compared to state-of-the-art methods on four datasets of 3D vascular segmentation tasks.
  arXiv  Detail & Related papers  (2023-06-02T01:52:35Z)
• The Topology-Overlap Trade-Off in Retinal Arteriole-Venule Segmentation [0.0]
  convolutional neural networks can achieve high overlap between predictions and expert annotations. We show that our model is able to produce results on par with state-of-the-art from the point of view of overlap.
  arXiv  Detail & Related papers  (2023-03-31T13:01:05Z)
• Automated Coronary Arteries Labeling Via Geometric Deep Learning [13.515293812745343]
  We propose an intuitive graph representation method, well suited to use with 3D coordinate data obtained from angiography scans. We subsequently seek to analyze subject-specific graphs using geometric deep learning. The proposed models leverage expert annotated labels from 141 patients to learn representations of each coronary segment, while capturing the effects of anatomical variability within the training data.
  arXiv  Detail & Related papers  (2022-12-01T09:31:08Z)
• Few-shot Medical Image Segmentation using a Global Correlation Network with Discriminative Embedding [60.89561661441736]
  We propose a novel method for few-shot medical image segmentation. We construct our few-shot image segmentor using a deep convolutional network trained episodically. We enhance discriminability of deep embedding to encourage clustering of the feature domains of the same class.
  arXiv  Detail & Related papers  (2020-12-10T04:01:07Z)
• A Teacher-Student Framework for Semi-supervised Medical Image Segmentation From Mixed Supervision [62.4773770041279]
  We develop a semi-supervised learning framework based on a teacher-student fashion for organ and lesion segmentation. We show our model is robust to the quality of bounding box and achieves comparable performance compared with full-supervised learning methods.
  arXiv  Detail & Related papers  (2020-10-23T07:58:20Z)
• Structured Landmark Detection via Topology-Adapting Deep Graph Learning [75.20602712947016]
  We present a new topology-adapting deep graph learning approach for accurate anatomical facial and medical landmark detection. The proposed method constructs graph signals leveraging both local image features and global shape features. Experiments are conducted on three public facial image datasets (WFLW, 300W, and COFW-68) as well as three real-world X-ray medical datasets (Cephalometric (public), Hand and Pelvis)
  arXiv  Detail & Related papers  (2020-04-17T11:55:03Z)

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.