The document discusses statistical analysis of wavelet coefficients of thermographs for characterizing breast cancer. It presents a wavelet-based technique to detect breast cancer in thermographs. Haar, biorthogonal, and reverse biorthogonal wavelets are analyzed and it is found that Haar wavelets provide better results in representing the temperature variations in cancer-affected regions. The methodology involves applying discrete wavelet transforms to segmented thermographs and calculating statistical measures like mean and standard deviation of the approximation and detail coefficients. The absolute difference between corresponding left and right segments is used to detect abnormal regions indicative of cancer.