Directional wavelet based features for colonic polyp classification.

•We test several wavelet based approaches for the classification of colon polyps.•A total of 11 differernt endoscopic polyp databases is used for our experiments.•The best results are achieved by extracting Weibull features from the subbands.•We propose three wavelet based feature extraction approaches using Weibull features.•Some of the methods improve the state of the art in classifying colonic polyps.