Malignancy Grade Identification and Classification of Brain MR Images with New 2D Co-occurrence Matrix and Wavelet Transformation

Malignancy is one of such terms in medical science that always requires quick attention. The proper identification and classification of such malignant cells were always considered as the challenging task. Moreover, with the use of computer assisted automation systems, identification of grading becomes easier but it requires a strong running algorithm for maintaining trust on results. This paper proposed a new algorithm in machine learning environment that fetches hidden patterns from the input MR images and found some relevant information about malignant cells. The proposed algorithm is the 2D co-occurrence matrix that uses pixel information for creating sample space. In addition of this 2D wavelet transformation was used to reduce the input image dimensions and fetching spectral information. The collective use of the DWT with proposed algorithm fetches better feature information about malignancy in brain MR images. The experimental result shows that the proposed approach gives better classification accuracy and performs well as compared with existing methods.