Classification of Findings with Localized Lesions in Fundoscopic Images using a Regionally Guided CNN
This work addresses the need for better automated diagnosis tools in ophthalmology by enhancing lesion localization, though it is incremental as it builds on existing CNN methods.
The paper tackled the problem of classifying and localizing lesions in fundoscopic images by proposing a regionally guided CNN with a new loss function, achieving up to 4.1% higher AUROC and improved localization compared to widely used networks.
Fundoscopic images are often investigated by ophthalmologists to spot abnormal lesions to make diagnoses. Recent successes of convolutional neural networks are confined to diagnoses of few diseases without proper localization of lesion. In this paper, we propose an efficient annotation method for localizing lesions and a CNN architecture that can classify an individual finding and localize the lesions at the same time. Also, we introduce a new loss function to guide the network to learn meaningful patterns with the guidance of the regional annotations. In experiments, we demonstrate that our network performed better than the widely used network and the guidance loss helps achieve higher AUROC up to 4.1% and superior localization capability.