Colorectal polyp detection is important in preventing cancer. Structural entropy can detect different structures of distributions, such as image pixel brightness. Wavelet analysis can help in separating large-scale and fine resolution behaviour. In the method presented in this paper, the colonoscopy images are separated into segments, and a classification scheme is built in order to determine, whether there is a polyp part in the image segment or not. Without wavelet analysis edge density and structural entropy can be a basis of fuzzy classification for the polyp content of only good quality colonoscopy images, and still has about 10 percent false classification. In this contribution the effect of wavelet analysis on the classification scheme is studied.