CVMar 8, 2023

Medical Waste Sorting: a computer vision approach for assisted primary sorting

arXiv:2303.04720v27 citationsh-index: 19
AI Analysis

This addresses the high-cost management of medical waste for hospitals and clinics, though it appears incremental as it applies existing computer vision methods to a new domain.

The paper tackles the problem of sorting hazardous medical waste to recover valuable materials by proposing a computer vision approach for assisted primary sorting, achieving 100% accuracy on a representative dataset and demonstrating good generalization on a new dataset.

Medical waste, i.e. waste produced during medical activities in hospitals, clinics and laboratories, represents hazardous waste whose management involves special care and high costs. However, this kind of waste contains a significant fraction of highly valued materials that can enter a circular economy process. To this end, in this paper, we propose a computer vision approach for assisting in the primary sorting of medical waste. The feasibility of our approach is demonstrated on a representative dataset we collected and made available to the community, with which we have trained a model that achieves 100\% accuracy, and a new dataset on which the trained model exhibits good generalization.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes