A NOVEL APPROACH FOR DETECTING BREAST TUMOR BOUNDARIES USING SEGMENTATION TECHNIQUE

Abstract

Mammography is the best and solid technique for the early determination of the bosom tumors through screening and exact identification of masses, microcalcifications and structural bends. Bosom disease is the primary driving reason for death for the lady in the world. It is seen that early identification of danger can help in the analysis of the infection for a lady and it can help emphatically to upgrade the anticipation of survival. For the discovery of bosom malignant growth, different methods are utilized in which mammography is the most encouraging system and utilized by radiologist now and again. Mammogram pictures are more often than not of low differentiation and uproarious. In bosom mammography, bright districts represent cancer. In some mammogram images, malignant tissues and ordinary thick tissues both might be available. To differentiate among threatening and ordinary thick tissues is beyond the realm of imagination just through applying thresholding. Understanding the information of mass regions of cancerous lesions in a mammogram is fundamental and distinguishes the tumour and its division. Along these lines, the discovery of injuries in mammogram pictures turns into a functioning exploration zone. Numerous procedures including PC helped location frameworks and power-based strategies were presented for bosom malignant growth division in mammogram pictures. The proposed approach expresses a method for segmenting the tumour boundaries from mammogram/MRI breast cancer images. This method is proposed to obtain an optimal segmented region. The first step begins with the process of finding the area which represents the tumour by using the proposed method. To achieve this, the regions of all the pixel present in the entire image is grown.

Keywords: Breast Tumor, Segmentation, micro classification, MRI image.