Enhancing Automatic Placenta Analysis through Distributional Feature Recomposition in Vision-Language Contrastive Learning
Research Poster Engineering 2025 Graduate ExhibitionPresentation by Max Mehta
Exhibition Number 1
Abstract
The placenta is a valuable organ that can aid in understanding adverse events during pregnancy and predicting issues post-birth. Manual pathological examination and report generation, however, are laborious and resource-intensive. Limitations in diagnostic accuracy and model efficiency have impeded previous attempts to automate placenta analysis. This study presents a novel framework for the automatic analysis of placenta images that aims to improve accuracy and efficiency. Building on previous vision-language contrastive learning (VLC) methods, we propose two enhancements, namely Pathology Report Feature Recomposition and Distributional Feature Recomposition, which increase representation robustness and mitigate feature suppression. In addition, we employ efficient neural networks as image encoders to achieve model compression and inference acceleration. Experiments validate that the proposed approach outperforms prior work in both performance and efficiency by significant margins. The benefits of our method, including enhanced efficacy and deployability, may have significant implications for reproductive healthcare, particularly in rural areas or low- and middle-income countries.
Importance
The infant mortality rate in low-income countries is over ten times higher than in high-income countries, which underlines the necessity for accessible healthcare. The placenta has discernable features that can serve as indicators of adverse pregnancy outcomes. In a clinical context, these adverse outcomes are often signaled by morphological changes in the placenta, identifiable through pathological analysis. Timely conducted placental pathology can reduce the risks of serious consequences of pregnancy-related infections. Unfortunately, traditional placenta pathology examination is time-consuming and resource-intensive, requiring specialized equipment and expertise. To overcome these challenges, our research explores the use of automatic placenta analysis tools that rely on photographic images. By enabling broader and more timely placental analysis, these tools could help reduce infant fatalities.
DEI Statement
Our research emphasizes diversity, equity, and inclusion in reproductive healthcare by ensuring broad accessibility to advanced placental analysis. Our work prioritizes accuracy and efficiency, optimizing the model for deployment on smaller mobile devices. This focus enhances feasibility in resource-limited settings and improves access to reproductive healthcare in underserved areas of the U.S., where healthcare disparities and policy changes may affect women's access to essential maternal health services. We also incorporate data from diverse sites, including Ghana, with varying demographics (age, race, cultural background), to ensure that our placental pathology models are representative of global populations, particularly in low- and middle-income countries (LMICs). This inclusivity improves diagnostic accuracy and supports equitable healthcare outcomes for women and newborns worldwide.