AI-assisted colonoscopy systems have been demonstrated to increase adenoma detection rates. We trained a YOLOv5 deep learning model combining bounding box predictions and object classification to detect polyps in colonoscopy videos.

We trained a deep learning model based on EfficientNetB0 that can diagnose exudative pharyngitis and deployed the model on a web application that can be used to augment the doctor’s diagnosis of exudative pharyngitis.

Automated labelling of radiology reports with natural language processing uses four main methods, namely: (1) rules-based text-matching, (2) conventional machine learning, (3) neural networks and (4) BERT. This paper details the necessary data preprocessing steps, reviews the 4 main methods for automated labelling and compares their performance.