# Evaluation of Blood Vessel Segmentation Methods on Hard-to-Detect Vascular Structures

arXiv: https://arxiv.org/abs/2406.13128 · HTML: https://mlsanity.com/paper/2406.13128/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 26 of its 65 references that are in the map.
- First posted on arXiv 19 Jun 2024.

## Builds on (most cited first, 26 of 26)

- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features](https://mlsanity.com/paper/1905.04899/) (2019)
- [UNet++: A Nested U-Net Architecture for Medical Image Segmentation](https://mlsanity.com/paper/1807.10165/) (2018)
- [Attention U-Net: Learning Where to Look for the Pancreas](https://mlsanity.com/paper/1804.03999/) (2018)
- [ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness](https://mlsanity.com/paper/1811.12231/) (2018)
- [Deep High-Resolution Representation Learning for Visual Recognition](https://mlsanity.com/paper/1908.07919/) (2019)
- [Road Extraction by Deep Residual U-Net](https://mlsanity.com/paper/1711.10684/) (2017)
- [UNet 3+: A Full-Scale Connected UNet for Medical Image Segmentation](https://mlsanity.com/paper/2004.08790/) (2020)
- [clDice -- A Novel Topology-Preserving Loss Function for Tubular Structure Segmentation](https://mlsanity.com/paper/2003.07311/) (2020)
- [ResUNet++: An Advanced Architecture for Medical Image Segmentation](https://mlsanity.com/paper/1911.07067/) (2019)
- [Topology-Preserving Deep Image Segmentation](https://mlsanity.com/paper/1906.05404/) (2019)
- [Metrics reloaded: Recommendations for image analysis validation](https://mlsanity.com/paper/2206.01653/) (2022)
- [Beyond the Pixel-Wise Loss for Topology-Aware Delineation](https://mlsanity.com/paper/1712.02190/) (2017)
- [Topology-Aware Segmentation Using Discrete Morse Theory](https://mlsanity.com/paper/2103.09992/) (2021)
- [SA-UNet: Spatial Attention U-Net for Retinal Vessel Segmentation](https://mlsanity.com/paper/2004.03696/) (2020)
- [Common Limitations of Image Processing Metrics: A Picture Story](https://mlsanity.com/paper/2104.05642/) (2021)
- [OCTA-500: A Retinal Dataset for Optical Coherence Tomography Angiography Study](https://mlsanity.com/paper/2012.07261/) (2020)
- [DCSAU-Net: A Deeper and More Compact Split-Attention U-Net for Medical Image Segmentation](https://mlsanity.com/paper/2202.00972/) (2022)
- [Attention Guided Network for Retinal Image Segmentation](https://mlsanity.com/paper/1907.12930/) (2019)
- [Topology-Aware Uncertainty for Image Segmentation](https://mlsanity.com/paper/2306.05671/) (2023)
- [A skeletonization algorithm for gradient-based optimization](https://mlsanity.com/paper/2309.02527/) (2023)
- [Human Treelike Tubular Structure Segmentation: A Comprehensive Review and Future Perspectives](https://mlsanity.com/paper/2207.11203/) (2022)
- [An Analysis of the Influence of Transfer Learning When Measuring the Tortuosity of Blood Vessels](https://mlsanity.com/paper/2111.10255/) (2021)
- [A new dataset for measuring the performance of blood vessel segmentation methods under distribution shifts](https://mlsanity.com/paper/2301.04517/) (2023)
- [TopoAL: An Adversarial Learning Approach for Topology-Aware Road Segmentation](https://mlsanity.com/paper/2007.09084/) (2020)

## Abstract (arXiv)

Due to the intricate structure of vascular trees, minor segmentation errors can significantly alter connectivity patterns and increase variability in extracted morphological properties. Global metrics such as the Dice coefficient, precision, and recall often overlook inaccuracies in specific regions of a sample. To address this, we define a Local Vessel Salience (LVS) index to quantify the difficulty of identifying specific vessel segments. This index is used to evaluate the performance of 16 segmentation methods across six widely used 2D datasets. The LVS index is calculated for each vessel pixel by comparing local vessel intensity against the surrounding background. We introduce a metric termed mean Low-Salience Recall (mLSR) to quantify how effectively algorithms recover hard-to-detect vessels. Furthermore, we propose a proof-of-concept data augmentation procedure guided by the LVS index aimed at improving neural network segmentation performance. Our findings demonstrate that segmentation performance strongly correlates with LVS, revealing systematic errors in vessels with low salience. The mLSR across all evaluated methods was significantly lower than standard recall values, with a typical performance decrease of 20 percentage points. In benchmark datasets such as DRIVE and OCTA-500, mLSR values for the top-performing methods were approximately 58%. The developed methodology provides a quantitative basis for the design of segmentation algorithms with improved sensitivity to hard-to-detect vessels and enhanced capabilities for preserving vascular connectivity.