Global Journal of Science Frontier Research, A: Physics and Space Science, Volume 22 Issue 1

(A) (B) (C) (D) Figure 9: Sample 6 shows, (A) the original image, (B) cluster 1, (C) cluster 2 and (D) cluster 3 after separating objects by colour using K-means clustering technique in L*a*b* colour space with the tumor region circled in yellow (A) (B) (C) (D) Figure 10: Sample 7 shows, (A) the original image, (B) cluster 1, (C) cluster 2 and (D) cluster 3 after separating objects by colour using K-means clustering technique in L*a*b* colour space with the tumor region circled in yellow Figure 11: Sample 8 shows, (A) the original image, (B) cluster 1, (C) cluster 2 and (D) cluster 3 after separating objects by colour using K-means clustering technique in L*a*b* colour space with no tumor region Figures 4 to 11 show the defect segmentation result of the breast with masses and lesions using the K- means clustering technique. After segmenting the input image into four clusters in figures 4 to 11, it was affirmative that the fourth cluster correctly segmented the tumor portion of the image[1]. From the empirical observations, it was observed that using 3 or 4 clusters yielded good segmentation results. Thus, in this experiment, the input images were partitioned into four segments as it also shows the detection result on an image mass while considering a different number of clusters for K-Mean clustering[4]. When the number of clusters is set to 2, one cluster contains the breast part while another one contains mass and background. If the number of clusters is increased to 3, the defective part is separated with a background. Hence, we further increased the number of clusters to 4. In Figures 4 to 11, samples 1 to 8, the segmentation result was better for 4 clusters than 3 clusters because the area of masses and lesions in the breast is less and the Segmentation of Cancerous Mammography using MATLAB 1 Year 2022 49 © 2022 Global Journals Global Journal of Science Frontier Research Volume XXII Issue ersion I VI ( A )