Scholarlink

DNA-based data storage represents a cutting-edge method for the preservation of digital information. This approach encompasses standard processes such as data encoding, synthesis, storage, sequencing, and decoding. Taking advantage of distinctive channel characteristics specific to DNA and multi-read functionalities, we introduce a novel decider mechanism designed to assess the accuracy of data recovery amidst unknown noise interference. Forevermore, we present a decision algorithm engineered to autonomously evaluate the integrity of retrieved data, thus achieving high precision in DNA storage. We utilize edit distance to quantify the similarity between DNA strands, which constitutes a critical element of our decider mechanism for assessing the quality of recovered images. Experimental evaluations affirm the decider’s capacity to effectively differentiate between high-quality and compromised data. The comparative analysis highlights the decision algorithm’s superior performance over conventional methodologies. In contexts utilizing neural network-based strategies, our algorithm enhances the Peak Signal-to-Noise Ratio (PSNR) by 0.7309 dB and the Structural Similarity Index (SSIM) by 0.0179; in traditional settings, it improves PSNR by 2.1396 dB and SSIM by 0.0662.

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

Deep Decision Algorithm for DNA Image Storage: Enhancing Accuracy with Edit Distance-Based Quality Assessment

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Wenfeng Wu,
  • Luping Xiang,
  • Qiang Liu,
  • Kun Yang

摘要

DNA-based data storage represents a cutting-edge method for the preservation of digital information. This approach encompasses standard processes such as data encoding, synthesis, storage, sequencing, and decoding. Taking advantage of distinctive channel characteristics specific to DNA and multi-read functionalities, we introduce a novel decider mechanism designed to assess the accuracy of data recovery amidst unknown noise interference. Forevermore, we present a decision algorithm engineered to autonomously evaluate the integrity of retrieved data, thus achieving high precision in DNA storage. We utilize edit distance to quantify the similarity between DNA strands, which constitutes a critical element of our decider mechanism for assessing the quality of recovered images. Experimental evaluations affirm the decider’s capacity to effectively differentiate between high-quality and compromised data. The comparative analysis highlights the decision algorithm’s superior performance over conventional methodologies. In contexts utilizing neural network-based strategies, our algorithm enhances the Peak Signal-to-Noise Ratio (PSNR) by 0.7309 dB and the Structural Similarity Index (SSIM) by 0.0179; in traditional settings, it improves PSNR by 2.1396 dB and SSIM by 0.0662.