Incremental Cross-Domain Adaptation for Robust Retinopathy Screening via Bayesian Deep Learning
This work addresses the challenge of data scarcity in medical imaging for retinopathy screening, offering a practical solution for healthcare applications, though it is incremental in its approach.
The paper tackles the problem of requiring large annotated datasets for retinopathy screening by introducing an incremental cross-domain adaptation instrument that enables deep classification models to learn from few-shot training, achieving an overall accuracy of 0.9826 and F1 score of 0.9846 on six public datasets.
Retinopathy represents a group of retinal diseases that, if not treated timely, can cause severe visual impairments or even blindness. Many researchers have developed autonomous systems to recognize retinopathy via fundus and optical coherence tomography (OCT) imagery. However, most of these frameworks employ conventional transfer learning and fine-tuning approaches, requiring a decent amount of well-annotated training data to produce accurate diagnostic performance. This paper presents a novel incremental cross-domain adaptation instrument that allows any deep classification model to progressively learn abnormal retinal pathologies in OCT and fundus imagery via few-shot training. Furthermore, unlike its competitors, the proposed instrument is driven via a Bayesian multi-objective function that not only enforces the candidate classification network to retain its prior learned knowledge during incremental training but also ensures that the network understands the structural and semantic relationships between previously learned pathologies and newly added disease categories to effectively recognize them at the inference stage. The proposed framework, evaluated on six public datasets acquired with three different scanners to screen thirteen retinal pathologies, outperforms the state-of-the-art competitors by achieving an overall accuracy and F1 score of 0.9826 and 0.9846, respectively.