Optical coherence tomography (OCT) is an imaging technique that provides fast, noninvasive data collection process to reconstruct three‐dimensional structures of epithelial tissue. Our full-field OCT system is able to capture images in sub-micron resolution so that all the cellular-levle information can be recorded. We integrated the FF-OCT system and a CNN-based classifier to work on the detection of squamous cell carcinoma. It achieved over 80% classification accuracy. This is made possible partly by the employment of an FF-OCT imaging system at sub-micron resolution and partly by the employment of a bottom-up feature extraction mechanism built into the classifier. This work has been published in Journal of Biophotonics.