AnnotAid: AI-Driven Data Annotation Tool for Histology Images

Peter Škríba and Adam Bublavý

Supervisor(s): Prof. Vanda Benešová

Slovak Technical University

Abstract: Automatic processing of digital histology images can greatly benefit from the utilization of deep learning methods. The development of such methods requires large amounts of annotated histological images. However, currently available annotation tools often have very poor usability, resulting in ineffective annotation processes. We aim to address the urgent need for a simplified approach to annotating histopathology images, a task that is crucial for advances in automated diagnosis and analysis. By combining our expertise, we strive to develop a user-friendly annotation tool integrated with state-of-the-art deep learning techniques. This tool is designed to alleviate the burden on pathologists during the annotation process by leveraging artificial intelligence models adapted to the various challenges in the field, such as the Nottingham Grading System of breast cancer. Through a comprehensive analysis of breast cancer and existing annotation tools, we propose a solution in the form of a multiplatform annotation tool powered by AI, developed in close cooperation with medical domain experts. By combining our knowledge and resources, we aim to bridge the gap between manual annotation processes and the potential of AI-based solutions, which will ultimately improve patient outcomes and advance medical research in breast cancer diagnosis. The annotation tool is available at: annotaid.com.
Keywords: Computer Vision, Design, Human-Computer Interaction, Image Processing
Full text:
Year: 2024