2026-10-11T15:02:58Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33867041 2026-09-16T17:46:02Z category_146 category_12 category_272 category_873 category_734 category_931 category_61 category_132 portal_5 item_type_3 month_year_09_2026

<p>Ablation study results of probability parameters.</p> Lili Fan (283203) Zhanglin Yang (12516397) Fan Guo (65416) Ping Fang (109406) Space Science Cell Biology Environmental Sciences not elsewhere classified Chemical Sciences not elsewhere classified Biological Sciences not elsewhere classified Information Systems not elsewhere classified Developmental Biology Infectious Diseases world industrial manufacturing two prominent limitations single hierarchical level representative recent methods made initial progress level amplitude perturbations insufficient feature diversity hierarchical dual enhancement feature extraction restricted surface defect segmentation preserving frequency perturbation annotated defect samples shot segmentation paradigms annotated samples precision segmentation surface defects xlink "> thereby mitigating shot settings severely challenged scale defects representational capacity model prone limited supervision limited number learning based extreme scarcity extensive experiments although meta achieve high ablation studies 4i demonstrate <div><p>In real-world industrial manufacturing, surface defect segmentation is severely challenged by the extreme scarcity of annotated defect samples, which renders exhaustively supervised approaches impractical. Although meta-learning based few-shot segmentation paradigms have made initial progress, existing methods still suffer from two prominent limitations. First, the limited number of support set samples leads to insufficient feature diversity, rendering the model prone to overfitting during training. Second, feature extraction restricted to a single hierarchical level and lacking contextual guidance constrains the representational capacity for complex and multi-scale defects. To address these issues, a novel few-shot surface defect segmentation method termed FH-Net is proposed, which aims to achieve high-precision segmentation of previously unseen categories using only a limited number of annotated samples. Specifically, a Semantic-Preserving Frequency Perturbation (SPFP) module is designed, which enriches support set feature diversity through foreground- and global-level amplitude perturbations in the Fourier domain, thereby mitigating the risk of overfitting under limited supervision. Furthermore, a Hierarchical Context Aggregation Module (HCAM) is constructed, which strengthens multi-scale defect representation via parallel feature extraction branches coupled with an attention-driven adaptive weighting fusion mechanism. Extensive experiments on FSSD-12 and Surface Defects-4i demonstrate that FH-Net achieves competitive segmentation performance compared with representative recent methods under both the 1-shot and 5-shot settings. Ablation studies further confirm the contribution of each core module to the overall performance improvement.</p></div> 2026-09-16T17:45:45Z Dataset Dataset 10.1371/journal.pone.0358329.t007 https://figshare.com/articles/dataset/_p_Ablation_study_results_of_probability_parameters_p_/33867041 CC BY 4.0