COMPARISON OF U-NET AND MASK R-CNN FOR THEIR SEGMENTATION ACCURACY IN BRAIN TUMOR DETECTION

ABSTRACT

This study delves into a comprehensive comparison between two prominent deep learning architectures, namely U-Net and Mask R-CNN, focusing specifically on their segmentation accuracy in the challenging task of brain tumor detection. By undertaking such a comparative analysis, this project aims to assess the effectiveness of these models while also revealing the underlying intricacies that govern their effectiveness in medical image analysis. Such a comparative study allows for discerning the unique strengths and weaknesses inherent in each architecture, shedding light on their respective capabilities and limitations. The methodology employed in this study compares the segmentation accuracy of Mask R-CNN and U-Net for brain tumor detection using a standardized MRI dataset. The dataset was divided into training and test sets, with the training set used for model training and the test set for performance evaluation. Both models were assessed using Mean Intersection over Union (IoU) and precision to determine which provided superior accuracy. The comparison between Mask R-CNN and UNet for brain tumor segmentation highlighted key differences in their performance. After the segmentation process, Mask R-CNN demonstrated a precision value of 75% while that of U-Net was 67%. This therefore implies that the precision value of Mask R-CNN is 8% higher than that of UNet in identifying and segmenting tumor boundaries. This comparison of both models displays the capability of U-Net and Mask R-CNN in brain tumor segmentataion and therefore in the field of image segmentation.