A.N. Bondarenko and A.V. Katsuk (Russia)
wavelet , multifractal dimensions, medical image recogni tion.
We present several methods for automatically detecting cervical cancer in digital color images. These approaches need for extracting feature vectors from biomedical images with cervical cancer. The proposed methods to cervical cancer detection have been evaluated on a database con taining more than 1200 images. A combination of multifractal spectrum and wavelet analyssis are used to extract object features that would in dicate the presence of a cancer cells.The neural net used to classification biomedical images, after statistical analysis of feature vectors. Our experiments have demonstrated the potential of combination of multifractal and wavelet analy sis.
Important Links:
Go Back
