Deep Learning Medical Image Segmentation Methods: A Survey

Document Type : Original Article

Authors

1 Computer Science Department, Faculty of Computers and Artificial Intelligence, Helwan University, Cairo, Egypt

2 School of Information Technology and Computer Science ITCS, Nile University, Giza, Egypt

Abstract

Medical image segmentation is essential for detecting and localizing tumors in medical image analysis. Image segmentation involves the identification of anatomical structures in images. Medical image segmentation starts with manual segmentation using Atlas methods, then auto-segmentation, facilitated by deep learning algorithms. Deep learning-based medical image segmentation retains a significant pledge in reducing treatment planning, radiation-related toxicities, and side effects. This study provides a complete overview of deep-learning medical image segmentation models. We review various deep-learning models and architectures applied to medical image segmentation, including fully convolutional networks, U-Net, and attention-based models. This literature review discusses using different loss functions, data augmentation techniques, and transfer learning in deep learning-based medical image segmentation and several types of medical image modality. Evaluation analysis encloses benchmark datasets for human body organs such as the brain, lungs, chest, and liver. Finally, we summarize the challenges and future directions of deep learning for medical image segmentation.
 

Highlights

 

 

Keywords

Full Text

 

 

 

 

References

 
 

Files

Share

How to cite

Statistics