Penyakit barah dada adalah satu penyakit serious yang sering melanda wanita berusia 40 tahun dan ke atas. Untuk mengesan pertumbuhan merbahaya dalam tisu buah dada, imej mamogram digunakan. Kadang kala ianya susah untuk mengenalpasti tumor dalam buah dada kerana terdapat halangan dalam imej mamogram. Sebab itu, objektif-objektif projek ini adalah untuk menjelajah teknik-teknik pemprosessan imej yang dapat diaplikasikan pada imej mamogram untuk meningkatkan kualitinya dan mengaplikasikan algoritma sekmentasi watershed untuk mengsekmentasi tumor. Teknik-teknik pemprosessan imej terpilih dan algoritma sekmentasi watershed kemudian diimplimentasi sebagai program C++ dengan menggunakan Borland C++ Builder 6. Daripada keputusan perbandingan dengan Matlab Image Processing Toolbox, program C++ itu dapat mengimplimenkan semua teknik-teknik pemprosesan imej terpilih kecuali algoritma sekmentasi watershed. Secara kesimpulannya, kebanyakkan objektif projek telah dicapai. Namun, garis-garis watershed terputus-putus dan sekmentasi lebihan sering berlaku selepas mengaplikasi sekmentasi watershed dan ini akan dijadikan subjek penyelidikan pada masa depan untuk mengatasinya.
_____________________________________________________________________________________
Breast cancer is a serious illness that tends to affect women of age 40 and above. As a way to detect abnormalities in breast tissue, mammogram images are used. The problem now is that sometimes it is difficult to tell whether or not there are any tumor masses in the breasts due to obstructions in the mammogram images. Hence, the objectives of this project are to explore suitable image processing techniques that can be applied to mammogram images to enhance its quality and to apply watershed segmentation algorithm to mammogram images to segment tumor masses. The selected image processing techniques and watershed segmentation algorithm were then implemented as a C++ program developed from Borland C++ Builder 6. From the result of comparisons with the Matlab Image Processing Toolbox, the C++ program was able implement all selected image processing techniques successfully except the watershed segmentation algorithm. In conclusion, most of the objectives of the project were achieved. However, discontinuous watershed lines and over segmentation tends to occur after applying watershed segmentation and this shall be the subject of future research to overcome it.
Exploring image processing techniques for mammogram images_Choong Huai Chyn_E3_2009_875004449_00003095271_NI