Authors
Verma, S., Malusare, A., Wang, M., Wang, L., Mahapatra, A., English, A. L., Cox, A. D., Broman, M., de Brot, S., Burcham, G., Knapp, D., Dhawan, D., Sola, M., Aggarwal, V., Grama, A., Lanman, N. A.
Abstract
Digital pathology has grown exponentially in recent years, which has created an urgent need for methods to analyze large histopathological image slides. We present an open-source semi-supervised deep learning framework to efficiently analyze and annotate whole slide images in digital pathology. Our software "AnnotateAnyCell" is an intelligent interactive web-based interface, which facilitates exploration and labeling of cellular structures using active contrastive learning, and dimensionality reduction techniques. We integrate an active learning mechanism that continually improves the predictive model based on new annotations, showing improvements in classification accuracy over iterations of human-in-the-loop feedback learning. AnnotateAnyCell aims to enable rapid annotation of large-scale histopathological datasets by assisting experts with quantitative and contextual information at a cellular level leading to efficient and accurate labeling of cells. This framework has the potential to accelerate cancer biology and personalized medicine research by dramatically reducing the time and effort required for cellular analysis in tissue samples.
Preprint server:
bioRxiv
The authors list and abstract were imported from bioRxiv on 05 Nov 2025.
Advertisement
Stats
- Recommendations n/a n/a positive of 0 vote(s)
- Views 50
- Comments 0