Search papers, labs, and topics across Lattice.
LUMOS, a new semi-supervised learning framework, addresses the challenges of limited annotations and varying label granularities in OCT retinal layer segmentation by using a Dual-Decoder Network with Hierarchical Prompting Strategy (DDN-HPS) to reduce pseudo-label noise. The framework also incorporates Reliable Progressive Multi-granularity Learning (RPML) with region-level reliability weighting to ensure stable cross-granularity alignment during training. Results across six OCT datasets show LUMOS significantly outperforms existing methods, demonstrating strong cross-domain and cross-granularity generalization.
Achieve state-of-the-art OCT retinal layer segmentation by reliably aligning data across varying annotation granularities, even with limited labeled data.
Optical Coherence Tomography (OCT) layer segmentation faces challenges due to annotation scarcity and heterogeneous label granularities across datasets. While semi-supervised learning helps alleviate label scarcity, existing methods typically assume a fixed granularity, failing to fully exploit cross-granularity supervision. This paper presents LUMOS, a semi-supervised universal OCT retinal layer segmentation framework based on a Dual-Decoder Network with a Hierarchical Prompting Strategy (DDN-HPS) and Reliable Progressive Multi-granularity Learning (RPML). DDN-HPS combines a dual-branch architecture with a multi-granularity prompting strategy to effectively suppress pseudo-label noise propagation. Meanwhile, RPML introduces region-level reliability weighing and a progressive training approach that guides the model from easier to more difficult tasks, ensuring the reliable selection of cross-granularity consistency targets, thereby achieving stable cross-granularity alignment. Experiments on six OCT datasets demonstrate that LUMOS largely outperforms existing methods and exhibits exceptional cross-domain and cross-granularity generalization capability.
