WILLIE: A Unified Framework and Benchmark for Wound Classification, Segmentation, and Localization

2026-10-04

AI Summary

This study addresses the fragmentation of models for classification, segmentation, and localization tasks in chronic wound assessment by proposing a unified multi-task learning framework alongside a standardized benchmark. Methodologically, through data integration and task-compatibility design, the framework enables collaborative multi-task analysis within a single forward pass. Furthermore, it reveals that a segmentation-based localization paradigm outperforms dedicated object detectors. Experimental results demonstrate that the proposed framework achieves a classification accuracy of 91.88%, a segmentation Dice coefficient of 91.41%, and a localization AP@0.5 of 96.23%. These findings establish the framework as an efficient and unified solution for intelligent clinical wound assessment.

Abstract

Chronic wound management affects over 8.2 million patients in the United States and imposes substantial clinical and economic burden. Clinical wound assessment commonly involves three coupled tasks: identifying wound type, delineating wound boundaries, and localizing the wound region for measurement and monitoring. Despite this clinical coupling, existing machine learning approaches typically address wound classification, segmentation, and localization using separate models. We present WILLIE, a unified framework and benchmark for wound classification, segmentation and localization that enables systematic evaluation of multi-task wound analysis under a common protocol. WILLIE harmonizes three public wound datasets into a shared benchmark and compares unified models across three scaling configurations against 10 single-task baselines. The best model achieves 91.88% classification accuracy, 91.41% Dice, and 96.23% AP@0.5 while producing all three outputs in a single forward pass. Beyond aggregate performance, our results show that segmentation-derived localization outperforms dedicated detection baselines in this benchmark, suggesting that box-based localization may be unnecessary for spatially coherent wound targets. Our findings highlight that effective multi-task learning in healthcare imaging depends not only on shared representations, but also on task formulation, compatibility, and benchmark design.

Problem

• wound classification
• wound segmentation
• wound localization
• multi-task learning
• benchmark

Innovation

• multi-task learning
• wound analysis
• unified framework
• benchmark
• segmentation-derived localization