Scholarlink

<p>Diatom taxonomy is integral to drowning site inference, aiding in event reconstruction and the differentiation between accidental, suicidal and homicidal drownings. Traditional methods, however, are labor-intensive, require significant expertise and depend on expensive instruments such as scanning electron microscopy (SEM) at high magnifications. To address these limitations, this pilot study introduces a simplified diatom classification method based on three morphological traits: symmetry, outline and pattern. This approach enables analysis using routine optical microscopy with magnifications below 400×. By applying this method, diatom assemblages from four distinct locations within a specific district were examined and a regional database for drowning site inference was established. Statistical techniques, including Mahalanobis distances and principal component analysis (PCA), were employed to infer drowning sites, achieving over 90% accuracy in both in vitro and in vivo models. This method demonstrates the potential for accurate drowning site inference using accessible tools and minimal taxonomic expertise. Nonetheless, further validation in continuous water systems and diverse geographic regions is essential to enhance its reliability and generalizability.</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

A simple optical microscopy-based diatom classification approach for forensic drowning site inference: a pilot study

  • 期刊论文

  • 发表日期：

    2025年6月5日

查看全文
  • Ao Ma,
  • Min Chen,
  • Qihan Yu,
  • Liqin Chen,
  • Ping Huang,
  • Ji Zhang

摘要

Diatom taxonomy is integral to drowning site inference, aiding in event reconstruction and the differentiation between accidental, suicidal and homicidal drownings. Traditional methods, however, are labor-intensive, require significant expertise and depend on expensive instruments such as scanning electron microscopy (SEM) at high magnifications. To address these limitations, this pilot study introduces a simplified diatom classification method based on three morphological traits: symmetry, outline and pattern. This approach enables analysis using routine optical microscopy with magnifications below 400×. By applying this method, diatom assemblages from four distinct locations within a specific district were examined and a regional database for drowning site inference was established. Statistical techniques, including Mahalanobis distances and principal component analysis (PCA), were employed to infer drowning sites, achieving over 90% accuracy in both in vitro and in vivo models. This method demonstrates the potential for accurate drowning site inference using accessible tools and minimal taxonomic expertise. Nonetheless, further validation in continuous water systems and diverse geographic regions is essential to enhance its reliability and generalizability.