This document presents a method for analyzing pancreas histological images to identify glucose intolerance using wavelet decomposition. 134 pancreas images from rats were used, with 56 normal and 78 pre-diabetic. Wavelet analysis was used to segment beta cell regions. Features like area ratios were extracted and classifiers like SVM, MLP, Naive Bayes were applied, achieving up to 84% accuracy. Future work could introduce more robust features and faster segmentation methods to improve classification.