Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

U-Net in Medical Image Segmentation: A Review of Its Applications Across Modalities

Fnu Neha, Deepshikha Bhati, Deepak Kumar Shukla, Sonavi Makarand Dalvi, Nikolaos Mantzou, Safa Shubbar

TL;DR

Medical image segmentation across X-ray, MRI, CT, and Ultrasound is pivotal for diagnosis and treatment planning. The paper surveys U-Net and its variants (U-Net++, U-Net 3+) and analyzes their architectural differences, modality applications, and performance implications. It highlights limitations such as data scarcity, noise and artifacts, and computational complexity, and proposes directions like model compression, synthetic data, multimodal transformers, and explainable AI. The review aims to guide researchers and clinicians in selecting appropriate architectures and datasets to improve clinical decision-making and patient outcomes.

Abstract

Medical imaging is essential in healthcare to provide key insights into patient anatomy and pathology, aiding in diagnosis and treatment. Non-invasive techniques such as X-ray, Magnetic Resonance Imaging (MRI), Computed Tomography (CT), and Ultrasound (US), capture detailed images of organs, tissues, and abnormalities. Effective analysis of these images requires precise segmentation to delineate regions of interest (ROI), such as organs or lesions. Traditional segmentation methods, relying on manual feature-extraction, are labor-intensive and vary across experts. Recent advancements in Artificial Intelligence (AI) and Deep Learning (DL), particularly convolutional models such as U-Net and its variants (U-Net++ and U-Net 3+), have transformed medical image segmentation (MIS) by automating the process and enhancing accuracy. These models enable efficient, precise pixel-wise classification across various imaging modalities, overcoming the limitations of manual segmentation. This review explores various medical imaging techniques, examines the U-Net architectures and their adaptations, and discusses their application across different modalities. It also identifies common challenges in MIS and proposes potential solutions.

U-Net in Medical Image Segmentation: A Review of Its Applications Across Modalities

TL;DR

Medical image segmentation across X-ray, MRI, CT, and Ultrasound is pivotal for diagnosis and treatment planning. The paper surveys U-Net and its variants (U-Net++, U-Net 3+) and analyzes their architectural differences, modality applications, and performance implications. It highlights limitations such as data scarcity, noise and artifacts, and computational complexity, and proposes directions like model compression, synthetic data, multimodal transformers, and explainable AI. The review aims to guide researchers and clinicians in selecting appropriate architectures and datasets to improve clinical decision-making and patient outcomes.

Abstract

Medical imaging is essential in healthcare to provide key insights into patient anatomy and pathology, aiding in diagnosis and treatment. Non-invasive techniques such as X-ray, Magnetic Resonance Imaging (MRI), Computed Tomography (CT), and Ultrasound (US), capture detailed images of organs, tissues, and abnormalities. Effective analysis of these images requires precise segmentation to delineate regions of interest (ROI), such as organs or lesions. Traditional segmentation methods, relying on manual feature-extraction, are labor-intensive and vary across experts. Recent advancements in Artificial Intelligence (AI) and Deep Learning (DL), particularly convolutional models such as U-Net and its variants (U-Net++ and U-Net 3+), have transformed medical image segmentation (MIS) by automating the process and enhancing accuracy. These models enable efficient, precise pixel-wise classification across various imaging modalities, overcoming the limitations of manual segmentation. This review explores various medical imaging techniques, examines the U-Net architectures and their adaptations, and discusses their application across different modalities. It also identifies common challenges in MIS and proposes potential solutions.

Paper Structure

This paper contains 44 sections, 8 equations, 9 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Research Methodology
  3. Research Objective
  4. Medical Imaging Modalities
  5. X-Ray
  6. Properties of X-rays
  7. Image Formation
  8. Magnetic Resonance Imaging (MRI)
  9. Types of MRI
  10. Computed Tomography (CT)
  11. Image Planes: Axial, Coronal, and Sagittal
  12. Types of CT Scans and Phases
  13. Ultrasound(US)
  14. Working Principle of Ultrasound
  15. Types of US Scans
  16. ...and 29 more sections

Figures (9)

  • Figure 1: X-Ray - a non-invasive diagnostic medical imaging modality taken from wiki:xxxwiki:xxx1wiki:xxx2wiki:xxx3wiki:xxx4wiki:xxx5
  • Figure 2: Brain MRI - a non-invasive high-resolution medical imaging modality taken from gu2023exploring Fig. 2(a) T1-weighted MRI, 2(b) T2-weighted MRI, 2(c) Fluid-Attenuated Inversion Recovery (FLAIR), 2(d) Diffusion Tensor Imaging (DTI), 2(e) Susceptibility Weighted Imaging (SWI), 2(f) Functional MRI (fMRI)
  • Figure 3: Kidney CT scan taken from heller2019kits19 - Fig. 3(a) axial, 3(b) coronal, and 3(c) sagittal plane
  • Figure 4: Liver CT scan taken from young2015benign. Fig. 4(a) Non-Contrast CT, 4(b) Contrast Enchanced CT - Arterial Phase, 4(c) Venous Phase, 4(d) Delayed Phase
  • Figure 5: US image of the fetus at 12 weeks of pregnancy in a sagittal scan
  • ...and 4 more figures