BRAIN TUMOR IDENTIFICATION AND CLASSIFICATION OF MRI IMAGES USING DEEP LEARNING TECHNIQUES
K. Banupriya, R Kavya
Abstract
The primary goal of medical imaging is to extract meaningful and accurate information from these images with the least error possible. A system to decide whether the brain has a tumor or not from the MR image using the combined technique of auto stacked encoder methodology. In the first stage, the input image is converted to greyscale using binary thresholding and the spots are detected. The set of features extracted are characterized by using the DNN algorithm, then the tumor recognition is done with fine-tuning. The most efficient and effective algorithms are discussed after studying several relevant research dissertations. Pre-processing brain images, segmenting them, feature extraction and detection of the tumor are the approaches in most researches. These techniques, limitations, and the advantages with further expansion are discussed extensively in the paper. Image Processing is a strategy that changes over the normal Image into advanced structure so as to upgrade the nature of an Image quality which gives valuable data. The different application, for example, report handling, entertainment world, clinical imaging, measurable investigations, distant detecting, and military relies upon the Image processing method.