Quantitative cell imagery in cancer pathology has progressed greatly in the last 30 years where the diagnosis is still critically reliant upon the analysis of biopsy samples which are usually analyzed by a trained pathologist who, by analyzing the biopsies under microscope, assesses the normality or malignancy of the samples submitted. The diagnosis is crucial for determining a course of treatment. Recently, a pioneering work by the LPI has concentrated on using multispectral imagery using up to 64 spectral bands. There is increasing evidence that the much wider area beyond the visible spectrum is an interesting medium to study from a pathological point of view. In particular, the mid-infrared (MIR) bands contain richer information content than the visible bands. This proposal will investigate the use of multispectral imagery in both visible and MIR bands. We intend to combine information from the multispectral bands of both the visible and MIR bands with view to understanding the biochemistry of colorectal cancer and hence develop a robust and accurate multispectral computer aided system for automatic colorectal cancer detection, classification and diagnosis using visible and non-visible spectra. This will be used by the pathologists to assist them in the early detection of colorectal cancer. In addition, the methods, software and hardware tools which will be produced should be very applicable to classification of other cancer types such as lung, breast, cervix etc.