ANALISIS CITRA MAMMOGRAFI DALAM IDENTIFIKASI KEBERADAAN MASS TUMOR PAYUDARA MENGGUNAKAN EKSTRAKSI CIRI
Jumriati · Hasanuddin University Repository · 2016
A research of analysis mammography image to identify the presence of breast tumor mass has been conducted using feature extraction. The data used consists of 15 mammae images: 8 masses and 7 non-masses. A computer program is made in matlab for data analysis by firstly doing cropping in the part of data which contain mass and non-mass. Then the data are pre-processed using a filter median to remove image noise. Furthermore, the process color segmentation uses Otsu Two separate object and background to get the average value (??). The obtained image is marked by white color (value 255) and black color (0). The next step conducted is edge detection process, using Sobel method to reduce the noise of image, so it could ridge lines information which is thick, thin and clipped, which are characterized by different colors in the image. The next is feature extraction in roundness shape, aspect ratio, and feature extraction texture with GLCM and the results are able to analyze roundness and texture of the image. The Results of color segmentation Otsu shows that the image of experiencing separation by color limit is 240, whereas the statistics on GLCM feature shows the relationship between ASM and the greatest entropy with the correlation was 0.764 (R?? = 0.764).