Baiker Martin, Pieterman René, Zoon Peter
Department of Microtraces, Netherlands Forensic Institute, Laan van Ypenburg 6, Den Haag 2497GB, The Netherlands.
Department of Microtraces, Netherlands Forensic Institute, Laan van Ypenburg 6, Den Haag 2497GB, The Netherlands.
Forensic Sci Int. 2015 Jun;251:40-9. doi: 10.1016/j.forsciint.2015.03.003. Epub 2015 Mar 18.
The traditional way of visual toolmark comparison includes subjective judgments. Automated methods using computers are a possibility to render a comparison more objective, but they require the statistical properties, like the similarity and variability, of toolmarks to be determined quantitatively. Several parameters, that play a role during toolmark creation, are statistically analyzed in this article. We determined the same toolmark and the different toolmark similarity as well as variability of known matching toolmarks created in wax and compared the results with the similarity and variability of known non-matching toolmarks. In addition we studied the influence of the substrate materials wax and lead and the angle of attack on toolmark similarity and variability. Furthermore, we present an approach to determine toolmark quality, defined as how well structural details are preserved in the toolmark, to assist toolmark examiners in deciding, which structural details are reliable in a mark. We studied the influence of the substrate material, the angle of attack and the depth of a toolmark on the quality. The results show that for known matching toolmarks, the variability is very low within a toolmark and between toolmarks in wax, given that the parameters angle of attack and depth are held constant. Geometrical details are reliably represented down to 10-50μm and toolmark similarity is clearly higher than known non-matching similarities. The comparison of wax and lead shows that wax is a good alternative as a substrate material for experimental toolmarks, capable of reliably representing structural details down to 10-25μm. For finer details, lead is a better choice but might alter the original state of a tool. With increasing angle of attack, toolmark variability increases and toolmark quality decreases. Therefore it is advantageous to push the tool instead of pulling during toolmark creation for angles of attack above ≈45°. The quality also decreases with increasing toolmark depth, but only up to ≈300μm. Therefore toolmarks should be created as shallow as possible in the substrate material.
传统的视觉工具痕迹比较方法包含主观判断。使用计算机的自动化方法有可能使比较更加客观，但这需要对工具痕迹的统计特性（如相似度和变异性）进行定量测定。本文对在工具痕迹形成过程中起作用的几个参数进行了统计分析。我们测定了在蜡中制作的已知匹配工具痕迹的相同工具痕迹、不同工具痕迹的相似度以及变异性，并将结果与已知不匹配工具痕迹的相似度和变异性进行了比较。此外，我们研究了基底材料蜡和铅以及攻角对工具痕迹相似度和变异性的影响。此外，我们提出了一种确定工具痕迹质量的方法，该质量定义为工具痕迹中结构细节的保留程度，以协助工具痕迹检验人员判断标记中哪些结构细节是可靠的。我们研究了基底材料、攻角和工具痕迹深度对质量的影响。结果表明，对于已知匹配的工具痕迹，在攻角和深度等参数保持不变的情况下，工具痕迹内部以及蜡中工具痕迹之间的变异性非常低。几何细节在10 - 50μm的范围内能够可靠呈现，并且工具痕迹的相似度明显高于已知不匹配的相似度。蜡和铅的比较表明，蜡作为实验性工具痕迹的基底材料是一个很好的选择，能够可靠地呈现低至10 - 25μm的结构细节。对于更精细的细节，铅是更好的选择，但可能会改变工具的原始状态。随着攻角的增加，工具痕迹变异性增加，工具痕迹质量下降。因此，对于攻角大于≈45°的情况，在工具痕迹制作过程中推工具而不是拉工具是有利的。质量也会随着工具痕迹深度的增加而下降，但仅在深度达到≈300μm之前如此。因此，应在基底材料中尽可能浅地制作工具痕迹。
