Scholarlink

<p>In the dicing machine cutting process, wafer alignment is a crucial part, and its error will directly affect the yield of components. Currently, traditional intelligent optimization algorithms have been widely used in the alignment task, but there is a problem of low accuracy. To address the above challenge, this study proposes a multi-strategy improved sparrow search algorithm based on a multi-scale residual channel attention network. First, mutual information is selected as the objective function for the proposed algorithm, and a multi-scale residual channel attention super-resolution network is proposed to accurately compute the mutual information. Second, the population is initialized using Tent-Cubic chaotic mapping to make the population distribution more uniform. Finally, an improved sine-cosine strategy as well as a hybrid strategy of Cauchy-Gaussian variation and variable spiral search is introduced to enhance the global search and local exploration capabilities of the algorithm. Several experiments show that the proposed method achieves higher alignment accuracy and stability.</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

Wafer sub-pixel alignment based on multi-scale residual channel attention network and multi-strategy improved sparrow search algorithm

  • 期刊论文

  • 发表日期：

    2025年8月12日

查看全文
  • Wanyong Liang,
  • Weilin Jia,
  • Jun Shi,
  • Zehui Xie,
  • Yangyang Hu,
  • Weifeng Cao

摘要

In the dicing machine cutting process, wafer alignment is a crucial part, and its error will directly affect the yield of components. Currently, traditional intelligent optimization algorithms have been widely used in the alignment task, but there is a problem of low accuracy. To address the above challenge, this study proposes a multi-strategy improved sparrow search algorithm based on a multi-scale residual channel attention network. First, mutual information is selected as the objective function for the proposed algorithm, and a multi-scale residual channel attention super-resolution network is proposed to accurately compute the mutual information. Second, the population is initialized using Tent-Cubic chaotic mapping to make the population distribution more uniform. Finally, an improved sine-cosine strategy as well as a hybrid strategy of Cauchy-Gaussian variation and variable spiral search is introduced to enhance the global search and local exploration capabilities of the algorithm. Several experiments show that the proposed method achieves higher alignment accuracy and stability.