Related Experiment Video
Updated: Nov 4, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.1K
Improved U-Net architecture with VGG-16 for brain tumor segmentation
Sourodip Ghosh1, Aunkit Chaki2, K C Santosh3
1KC's PAMI Research Lab - Computer Science, University of South Dakota, Vermillion, SD, 57069, USA.
Physical and Engineering Sciences in Medicine
|May 28, 2021
Summary
We developed an improved U-Net model with VGG-16 for segmenting brain MRI images and detecting tumors. This advanced model achieved higher pixel accuracy than standard U-Net, outperforming existing methods for neurological disease detection.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Neurology
Background:
- Automated segmentation of Brain MRI images is crucial for detecting neurological diseases.
- Accurate identification of tumor cells in brain scans aids diagnosis and treatment planning.
Purpose of the Study:
- To propose an improved U-Net architecture integrated with VGG-16 for enhanced brain MRI segmentation.
- To accurately identify tumor cells within brain MRI images.
- To compare the performance of the proposed model against a custom U-Net architecture.
Main Methods:
- Utilized the TCGA-LGG dataset comprising 3929 brain MRI images from the TCI archive.
- Implemented an improved U-Net model incorporating VGG-16 for feature extraction.
- Compared the improved U-Net against a custom-designed U-Net architecture.
Main Results:
- Achieved a pixel accuracy of 0.9975 with the improved U-Net architecture.
- The improved U-Net demonstrated superior performance compared to the basic U-Net (0.994 pixel accuracy).
- The proposed method outperformed common Convolutional Neural Network (CNN)-based state-of-the-art approaches.
Conclusions:
- The improved U-Net with VGG-16 offers superior performance for brain MRI segmentation and tumor cell identification.
- This approach shows significant potential for advancing the automated assessment of neurological disorders.
- The model's high accuracy supports its application in clinical settings for improved diagnostic capabilities.
