Fusion of colour contrasted images for early detection of oesophageal squamous cell dysplasia from endoscopic videos in real time

•One of the first systems for multi-class detection of early onset of oesophageal cancer based on multi-modal imaging modalities.•Fusion of human perception through the embedding of contrasted images enhanced by applying a human colour vision model.•Improving system generalization by rendering contrasted images under the same viewing conditions.•Overcoming small dataset by augmenting data using enhanced contrasted images.•Achieving state of the art results while maintaining the transparency of the system and being operative in real time.