Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Segmentation Strategies in Deep Learning for Prostate Cancer Diagnosis: A Comparative Study of Mamba, SAM, and YOLO

Ali Badiezadeh, Amin Malekmohammadi, Seyed Mostafa Mirhassani, Parisa Gifani, Majid Vafaeezadeh

TL;DR

The paper addresses the challenge of segmenting prostate cancer histopathology images by benchmarking three deep-learning approaches—Mamba-based architectures (notably H-vmunet), SAM, and YOLO—on two comprehensive datasets, Gleason 2019 and SICAPv2. The authors introduce the High-order Vision Mamba UNet (H-vmunet) and detail its architecture incorporating High-order Visual State Space modules and 2D-selective-scan operations to balance local detail with global context. Across both datasets, H-vmunet delivers the highest segmentation accuracy, with $DSC$ and related metrics outperforming SAM and YOLO, demonstrating robustness across different tissue patterns and scales. The findings underscore H-vmunet’s potential for clinical deployment in computer-aided diagnosis and highlight the need for robust validation and efficiency improvements for real-time integration.

Abstract

Accurate segmentation of prostate cancer histopathology images is crucial for diagnosis and treatment planning. This study presents a comparative analysis of three deep learning-based methods, Mamba, SAM, and YOLO, for segmenting prostate cancer histopathology images. We evaluated the performance of these models on two comprehensive datasets, Gleason 2019 and SICAPv2, using Dice score, precision, and recall metrics. Our results show that the High-order Vision Mamba UNet (H-vmunet) model outperforms the other two models, achieving the highest scores across all metrics on both datasets. The H-vmunet model's advanced architecture, which integrates high-order visual state spaces and 2D-selective-scan operations, enables efficient and sensitive lesion detection across different scales. Our study demonstrates the potential of the H-vmunet model for clinical applications and highlights the importance of robust validation and comparison of deep learning-based methods for medical image analysis. The findings of this study contribute to the development of accurate and reliable computer-aided diagnosis systems for prostate cancer. The code is available at http://github.com/alibdz/prostate-segmentation.

Segmentation Strategies in Deep Learning for Prostate Cancer Diagnosis: A Comparative Study of Mamba, SAM, and YOLO

TL;DR

The paper addresses the challenge of segmenting prostate cancer histopathology images by benchmarking three deep-learning approaches—Mamba-based architectures (notably H-vmunet), SAM, and YOLO—on two comprehensive datasets, Gleason 2019 and SICAPv2. The authors introduce the High-order Vision Mamba UNet (H-vmunet) and detail its architecture incorporating High-order Visual State Space modules and 2D-selective-scan operations to balance local detail with global context. Across both datasets, H-vmunet delivers the highest segmentation accuracy, with and related metrics outperforming SAM and YOLO, demonstrating robustness across different tissue patterns and scales. The findings underscore H-vmunet’s potential for clinical deployment in computer-aided diagnosis and highlight the need for robust validation and efficiency improvements for real-time integration.

Abstract

Accurate segmentation of prostate cancer histopathology images is crucial for diagnosis and treatment planning. This study presents a comparative analysis of three deep learning-based methods, Mamba, SAM, and YOLO, for segmenting prostate cancer histopathology images. We evaluated the performance of these models on two comprehensive datasets, Gleason 2019 and SICAPv2, using Dice score, precision, and recall metrics. Our results show that the High-order Vision Mamba UNet (H-vmunet) model outperforms the other two models, achieving the highest scores across all metrics on both datasets. The H-vmunet model's advanced architecture, which integrates high-order visual state spaces and 2D-selective-scan operations, enables efficient and sensitive lesion detection across different scales. Our study demonstrates the potential of the H-vmunet model for clinical applications and highlights the importance of robust validation and comparison of deep learning-based methods for medical image analysis. The findings of this study contribute to the development of accurate and reliable computer-aided diagnosis systems for prostate cancer. The code is available at http://github.com/alibdz/prostate-segmentation.
Paper Structure (12 sections, 6 equations, 5 figures, 2 tables)

This paper contains 12 sections, 6 equations, 5 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Materials
  3. Gleason dataset
  4. SICAPv2 dataset
  5. Methods
  6. Mamba
  7. H-vmunet method (Architecture)
  8. SAM
  9. YOLO
  10. Results
  11. Discussion
  12. Conclusion

Figures (5)

  • Figure 1: Examples of Gleason 2019 annotations. Left G3, middle G4, and right G5.
  • Figure 2: Segmentation map estimation on a sample image from SICAPv2 dataset. Left: original image, middle: provided segmentation map, right: processed segmentation map with Gleason grade (G4) annotation.
  • Figure 3: (Top) The H-vmunet model architecture overview. (Bottom) Multi-level and multi-scale information fusion module (SAB, CAB) wu2024h.
  • Figure 4: (Left) Architecture of the H-VSS module. (Middle) Overview of n-order 2D-selective-scan module. (Right) Architecture of Local-SS2D module wu2024h
  • Figure 5: The pipeline of SAM in huang2024segment