Enhanced Region Growing for Brain Tumor MR Image Segmentation

Area of Science:

• Medical Imaging
• Computer-Aided Diagnosis
• Neurosurgery

Background:

• Brain tumors are a significant cause of mortality in both children and adults.
• Accurate segmentation of brain tumors is crucial for diagnosis and treatment planning.
• Current region-growing algorithms often rely on manual or semi-manual seed point initialization, impacting segmentation accuracy.

Purpose of the Study:

• To develop an enhanced region-growing algorithm for automatic seed point initialization in brain tumor segmentation.
• To improve the accuracy and efficiency of brain tumor segmentation compared to existing methods.
• To validate the proposed algorithm's performance against state-of-the-art deep learning approaches.

Main Methods:

• An enhanced region-growing algorithm with automatic seed point initialization was proposed.
• A thresholding technique was used for skull stripping of brain images.
• Mean intensities of 8 image blocks were computed to select 5 blocks with maximum intensities as seed points.
• Region of Interest (ROI) segmentation was performed using the selected seed points.

Main Results:

• The proposed algorithm achieved a Dice Similarity Score (DSS) of 0.89 in the first experimental setup (15 images).
• DSS values of 0.90 and 0.80 were obtained in the second (12 images) and third (800 images) experimental setups, respectively.
• The average DSS across three experimental setups was 0.86, demonstrating robust performance.

Conclusions:

• The enhanced region-growing algorithm provides accurate and automatic brain tumor segmentation.
• The proposed method shows competitive performance against state-of-the-art deep learning algorithms.
• This approach offers a promising tool for improving brain tumor diagnosis and treatment planning.