Rozita Rastghalam - Department of Electrical Engineering, Najafabad Branch, Islamic Azad University, Isfahan,
Hossein Pourghassem - Department of Electrical Engineering Najafabad Branch, Islamic Azad University Isfahan,

Thermography is a noninvasive, non-radiating,fast, and painless imaging technique that is able to detect breasttumors much earlier than the traditional mammographymethods. In this paper, a novel breast cancer detection algorithmbased on spectral probable features is proposed to separatehealthy and pathological cases during breast cancer screening.Gray level co-occurrence matrix is made from image spectrum toobtain spectral co-occurrence feature. However, this feature isnot sufficient separately. To extract directional and probablefeatures from image spectrum, this matrix is optimized anddefined as a feature vector. By asymmetry analysis, left and rightbreast feature vectors are compared in which certainly, moresimilarity in these two vectors implies healthy breasts. Ourmethod is implemented on various breast thermograms that aregenerated by different thermography centers. Our algorithm isevaluated on different similarity measures such as Euclideandistance, correlation and chi-square. The obtained results showeffectiveness of our proposed algorithm.

Breast thermogram, Breast cancer detection, Asymmetric analysis, Image spectrum, Spectral probable feacture