Enhancing few-shot semantic segmentation via nested feature refinement and alignment

Few-shot semantic segmentation (FSS) faces challenges due to limited annotated data, leading to performance bottlenecks. However, in existing methods, spatial details suffer from irreversible loss at the early inference stage, forming a bottleneck that hinders performance breakthroughs. This paper introduces NERA-Net, a novel framework designed to address these challenges by focusing on intrinsic feature enhancement. NERA-Net employs a Nested Pyramid Refinement Module (NPRM) to recover spatial details and a Prototype-Pixel Semantic Alignment (PPSA) module to enhance feature discriminability. Additionally, the Generalized Parametric Rectified Linear Unit (GPReLU) is introduced to optimize gradient flow. Experiments on PASCAL-5 i and COCO-20 i benchmarks demonstrate that NERA-Net achieves state-of-the-art performance, with significant improvements in mIoU (e.g., +2.0 % and +2.3 % on PASCAL-5 i under 1-shot and 5-shot settings, respectively.). Our approach exhibits superior convergence efficiency and object perception capabilities, offering a promising solution for FSS tasks. The source code of this project is available at: https://github.com/ZUA-CV2023/NERA-Net .