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

and false negative breast images. The retrieved mammogram images are classified into left breast lesion, right breast lesion, and non-palpable left breast lesion. The images were retrieved in Joint Photographic Experts Group (JPEG)format. The pixels in the images were represented as an 8-bit word. The images were retrieved in Red, Green, Blue (RGB) format each with a pixel size of 500 × 500. The mammogram images were from diagnoses conducted by radiologists and clinicians using the Breast-i device. In taking the images, a patient undresses and in a darkened room, sitting slightly forward places the Breast-i light source on the inferior surface of (underneath) the breast. The patient views the superior aspect (top surface) of each breast, which should be uniformly bright except for typically a few darker lines corresponding to superficial blood vessels. The mammogram images are classified into three major cases: malignant, benign and normal. b) Developed Matlab algorithm for image analysis A flowchart of the method which is implemented in MATrix LABoratory (Matlab) application software (R2013a Matlab, Math Works Inc) is described in figure 3. The basic aim of the proposed approach is to segment colors automatically using the K-means clustering technique and L*a*b* colour space. The mammogram images were read into MATrix LABoratory (Matlab) application software (R2013a Matlab, MathWorks Inc) from a folder in which they were saved. The images were transformed from RGB to L*a*b* colour space. The L*a*b* colour space was used because it consisted of a luminosity layer 'L*' and two chromaticity layers in 'a*' and 'b*’. Using the L*a*b* colour space is computationally efficient because all of the colour informationis present in the 'a*' and 'b*' layers only[12]. The colors were then classified using K-Means clustering in the 'a*b*' space. To measure the difference between the two colors, the Euclidean distance metric was used.Each Pixel was labelled in the image from the Results of K-Means. For every pixel in the input, K- means computed an index corresponding to a cluster. Every pixel of the image was labelled with its cluster index , also the mean 'a' and 'b' value for each area was extracted. These values served as colour markers in the 'a*b' space. The index image was further processed to generate 3clusters based on colour information[12]. The pixels in the image was separated by colour using pixel labels, which resulted in different images based on the number ofclusters. The results of the nearest neighbor classification were displayed. The labelled matrix contained a colour label for each pixel in the mammogram image. The labelled matrix was then used to separate objects in the original image by colour. The index of each cluster containing the cancerous part of the mammogramwas determined because K-means doesnot return the same cluster index value every time but this was done using the center value of clusters, which contained the mean value of 'a*' and 'b*' for each cluster[12]. Segmentation of Cancerous Mammography using MATLAB © 2022 Global Journals 1 Year 2022 46 Global Journal of Science Frontier Research Volume XXII Issue ersion I VI ( A ) Classify colour using K-Means clustering in ‘a*b*’ space Generate image that segment the input image by colour into clusters Determine the cancerous cluster Read mammogram image Transform image from RGB to L*a*b* colour space Label each pixel in the image from the results of K-Means