Advanced Brain Tumor Detection Through Multimodal Image Fusion and Segmentation Techniques

In this study, we explore the effectiveness of CT and MRI image fusion for enhanced brain tumor detection. By integrating multiple pre-segmentation techniques such as K-means clustering, Canny edge detection, Sobel operator, region-based segmentation, and Otsu's method, we aim to improve the accuracy of tumor identification. Principal Component Analysis (PCA) is utilized to fuse the CT and MRI images, retaining significant features from both modalities. Post-fusion, Otsu's thresholding is applied to achieve optimal binary segmentation of the tumor regions. The final visualization is enhanced through contour detection and highlighting. Our results demonstrate that the combination of advanced pre-segmentation methods with PCA fusion and Otsu's thresholding significantly improves the detection and delineation of brain tumors. Performance metrics indicate superior accuracy and reliability, making this approach a valuable tool in medical imaging. The study demonstrates significant performance improvements in brain tumor detection through CT-MRI image fusion, with Otsu's method achieving a Dice Coefficient of 0.67 and a Jaccard Index of 0.58. These metrics indicate enhanced tumor delineation, making the proposed approach highly effective for accurate and reliable medical imaging.