Master of Science (MS)
Number of Pages
A statistical pattern recognition system for ultrasound medical images of prostatic tissue for cancer has been proposed. Using the autocorrelation method, the correct size of a statistical sliding window for feature extraction was defined. Known texture discrimination features have been tested for effectiveness. Another set of discriminating features, based on edge value distribution, Fourier power spectrum and wavelet transform has been derived and investigated. The set can be used as an input to a neural net classifier.
Cancer; Classification; Feature; Human; Images; Medical; Prostate; Prostate Cancer; Recognition; Selection; Tissue
Computer science; Diagnostic imaging; Oncology; Artificial intelligence
University of Nevada, Las Vegas
If you are the rightful copyright holder of this dissertation or thesis and wish to have the full text removed from Digital Scholarship@UNLV, please submit a request to firstname.lastname@example.org and include clear identification of the work, preferably with URL.
Tsarev, Valeri V, "Feature selection for classification of medical images of human tissue for cancer recognition" (1996). UNLV Retrospective Theses & Dissertations. 3223.