Scholarlink

<p>Accurate classification of porcelain relic fragments is essential for restoration. Traditional methods, which depend on visual traits like glaze color and patterns, perform poorly with visually similar fragments, especially when microscopic image samples are scarce. To solve these issues, we introduce InSwAV, which integrates involution-enhanced residual blocks for robust feature extraction and employs a swapped assignment mechanism. This mechanism aligns deep features with learnable cluster prototypes by enforcing consistency between differently augmented views of the same image. The model is optimized by minimizing a cross-entropy loss against cluster assignments, which reduces training time significantly. we constructed the Porcelain Relic Microscopic Images (PRMI) dataset of five classification, with data augmentation applied to enhance model robustness. Experimental results show that InSwAV achieves a classification accuracy of 96.2% on the porcelain relic microscopic image dataset, outperforming existing methods.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

InSwAV: involution enhanced feature clustering and swapped assignments for porcelain relic microscopic image classification

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月24日

查看全文
  • Yangyang Liu,
  • Jiatong Liu,
  • Xinda Liu,
  • Guohua Geng,
  • Zhan Li

摘要

Accurate classification of porcelain relic fragments is essential for restoration. Traditional methods, which depend on visual traits like glaze color and patterns, perform poorly with visually similar fragments, especially when microscopic image samples are scarce. To solve these issues, we introduce InSwAV, which integrates involution-enhanced residual blocks for robust feature extraction and employs a swapped assignment mechanism. This mechanism aligns deep features with learnable cluster prototypes by enforcing consistency between differently augmented views of the same image. The model is optimized by minimizing a cross-entropy loss against cluster assignments, which reduces training time significantly. we constructed the Porcelain Relic Microscopic Images (PRMI) dataset of five classification, with data augmentation applied to enhance model robustness. Experimental results show that InSwAV achieves a classification accuracy of 96.2% on the porcelain relic microscopic image dataset, outperforming existing methods.