Segmentation and Classification of Brain Tumor from MRI Images

Image processing and their capabilities can be utilized for classifying brain tumor from MRI images. The MRI images are first resized and pre-processed to improve intensity levels and noise reduction. Pre- processing includes techniques like histogram equalization and noise filtering, ensuring uniformity across the dataset for better accuracy in subsequent steps. For segmentation, the project employs CNN (Convolutional Neural Networks), which are highly effective in feature extraction and pattern recognition. The CNN model is trained to identify tumor regions within the MRI images with high precision. The features from CNN are used by KNN classifier and Support Vector Machine Classifier for classification. KNN is utilized for its simplicity and effectiveness in distinguishing between tumor and non-tumor regions based on proximity in feature space, while SVM is applied for its robustness in handling high-dimensional data and distinguishing between various tumor types. The combination of CNN for segmentation and KNN/SVM for classification ensures a comprehensive approach to tumor detection, providing a reliable, low-cost diagnostic tool for healthcare professionals. Segmentation metrics like accuracy , sensitivity, dice and precision are used for evaluating the robustness of the system.