Classification of Histopathology Images by Siamese and Triplet Network Embeddings

In computational pathology, one of the most investigated aspects concerns the analysis of patients' tissue images. This type of image is often called a Whole Slide Image (WSI) and is characterised by a high resolution that makes impractical, if not impossible, the classical analysis methods that process the image in a single step. Automatic and Deep Learning methods fit into this context in an attempt to provide automatic medical decision support. The paper presents a representation/classification schema showing three main contributions, firstly that, thanks to deep metric learning, it is possible to move the embeddings of the tissue images by minimizing the distance between elements of the same class and maximizing the distance between elements of different classes, and this makes easier the classification task; secondly it offers an effective way to visualize the clustered embeddings justifying the classification results; finally it presents a confidence score that indicates the accuracy of the classification result. Using four different datasets, we show how the proposed approach achieves results in line with and sometimes better than state-of-the-art methods, despite being a much simpler model. A further strength lies in the possibility of having a visualisation that makes the classification mechanism explicit. In addition, we provide a classification confidence score that can add useful information to the medical decision support system.