A Systematic Approach for MRI Brain Tumor Localization and Segmentation Using Deep Learning and Active Contouring | JoVE Visualize

A Systematic Approach for MRI Brain Tumor Localization and Segmentation Using Deep Learning and Active Contouring

Shanaka Ramesh Gunasekara¹, H N T K Kaldera¹, Maheshi B Dissanayake¹

¹Department of Electrical and Electronic Engineering, Faculty of Engineering, University of Peradeniya, Kandy 20400, Sri Lanka.

Journal of Healthcare Engineering|March 29, 2021

View abstract on PubMed

Summary

This study introduces a novel deep learning method for accurate tumor segmentation. The three-stage approach combines convolutional neural networks and the Chan-Vese algorithm, achieving high reliability in glioma and meningioma boundary detection.

Area of Science:

Medical Image Analysis
Artificial Intelligence in Oncology

Background:

Accurate tumor boundary annotation and segmentation are critical for effective tumor extraction.
Traditional edge detection methods often struggle with the complexities of medical image segmentation.

Purpose of the Study:

To develop and evaluate a robust threefold deep learning architecture for precise tumor segmentation.
To improve the accuracy and reliability of identifying tumor boundaries in medical imaging.

Main Methods:

A deep learning architecture combining a deep convolutional neural network (CNN) for classification and a region-based convolutional neural network (R-CNN) for localization.
Utilizing the Chan-Vese segmentation algorithm, an active contour model, for precise tumor boundary contouring.
Performance evaluation using metrics such as Dice Score, Rand Index (RI), Variation of Information (VOI), and others, comparing against expert annotations.

Main Results:

The proposed architecture achieved an average Dice Score of 0.92 for both glioma and meningioma segmentation.
High performance was also indicated by other metrics including RI (0.9936), VOI (0.0301), GCE (0.004), BDE (2.099), PSNR (77.076), and MAE (52.946).
These results demonstrate the high reliability and effectiveness of the deep learning approach.

Conclusions:

The presented threefold deep learning architecture offers a reliable solution for accurate tumor segmentation.
The integration of CNNs, R-CNN, and the Chan-Vese algorithm provides superior performance in delineating tumor boundaries.
This method holds significant potential for improving oncological diagnostics and treatment planning.