Semantic Segmentation of Reflectance Confocal Microscopy Mosaics of Pigmented Lesions Using Weak Labels
Feb 15, 2021·
,,,,,·
0 min read
Mara D'Alonzo
Alican Bozkurt
Christi Alessi-Fox
Melissa Gill
Dana H. Brooks
Milind Rajadhyaksha
Kivanc Kose
Abstract
Reflectance confocal microscopy (RCM) is a non-invasive imaging tool widely used for skin cancer screening. Automated analysis of RCM images requires pixel-level annotations, which are costly and time-consuming to obtain. In this work, we present a weakly supervised approach for semantic segmentation of cellular patterns in RCM mosaics of pigmented lesions. Our method leverages image-level labels to train a segmentation model, significantly reducing the annotation burden. We evaluate our approach on a dataset of RCM mosaics from melanocytic lesions and demonstrate competitive performance compared to fully supervised methods.
Type
Publication
In Scientific Reports 11(1), 3679 (2021)
Authors
AI Scientist
I am an AI Scientist at Paige AI. I did my Ph.D. with Jennifer Dy, Dana Brooks, and Jan-Willem van de Meent at Northeastern University. My main research interests are machine learning with emphasis on probabilistic programming, deep neural networks, and their applications in biomedical image processing. I am one of the developers of Probabilistic Torch, a library for deep generative models that extends PyTorch. I am also one of the maintainers of the PyTorch distributions module.