Spectral and Statistical Texture Parameters in Fuzzy Neural Network for Cancerous Colorectal Cells Images Classification

This paper presents a novel method which automatically detects differences in colon cell images, extracts the required texture data and then classifies the cells into normal or malignant. The images are captured by a CCD camera from a laboratory microscope slide. The new system is implemented by fuzzifying image texture descriptors from the Fourier power spectrum and greyscale statistical co-occurrence matrix and feeding it into a fuzzy neural network classifier to classify the images. The novel system has been evaluated using 116 cancers and 88 normal colon cells images and resulted in 96.435% classification accuracy. The novelty of the algorithm is that it is independent of the feature extraction procedure adopted and overcomes the sharpness of class characteristics associated with other classifiers.