Search papers, labs, and topics across Lattice.
This paper introduces IPSM-Bench, the largest annotated dataset for segmenting intermediate phases in zinc-based alloy microstructures, addressing significant challenges such as low contrast and heterogeneous morphologies. The authors develop SCoP-SAM, a novel segmentation method that utilizes spatial context priors to enhance the encoding-decoding process, leading to superior segmentation outcomes. Experimental results show that SCoP-SAM achieves state-of-the-art performance on IPSM-Bench and generalizes effectively to other alloy datasets, marking a significant advancement in microstructure analysis.
SCoP-SAM not only sets a new benchmark for zinc-alloy intermediate phase segmentation but also excels in generalizing across various alloy scenarios.
Zinc-based alloys are indispensable emerging absorbable metallic biomaterials, and their macroscopic performance is governed by microstructural characteristics. Intermediate phases-key microstructural constituents-are pivotal in regulating mechanical and functional properties. However, intermediate phase segmentation in zinc alloy microstructures faces formidable challenges: scarce annotated datasets, low contrast, difficulty detecting small targets, and heterogeneous morphologies. To this end, we construct IPSM-Bench, the largest high-quality dataset for zinc-alloy intermediate phase segmentation. Furthermore, we propose SCoP-SAM, a new Spatial Context Prior-guided SAM method that leverages the gradient structure and grayscale properties of intermediate phases to capture spatial context priors and incorporates them into the entire SAM encoding-decoding process, improving segmentation performance. Based on the proposed IPSM-Bench, we establish a new benchmark for intermediate phase segmentation to systematically evaluate state-of-the-art (SOTA) methods and advance research on zinc alloy microstructure analysis. Extensive experiments on IPSM-Bench and additional public alloy benchmarks demonstrate that our SCoP-SAM not only achieves SOTA performance for zinc-alloy intermediate phase segmentation but also generalizes remarkably well to other alloy scenarios.
