An objective inter-comparison of trash mark constellations generated by manual and automated detection methods

Capturing high‐resolution digital images for use in forensic document examination

“…Exemplar creation was required prior to execution of the digitization and semi-automated trash mark examination processes. Exemplars without any writing or printing were created and used for this study as proof of concept and a way of verifying that this method and the subsequent analysis to follow [3] five individual photocopies. The default settings for the multifunction office machines are provided in Table 1.…”

“…Both were evaluated and subsequently compared to obtain optimal images for later project tasks. The parameters discussed in the following sections in combination with one another produce the image later analyzed by the hybrid subject matter expert/machine method [3], and the results obtained compared to the ground truth manual examination were used to assess the results and draw the conclusions.…”

“…To bolster foundational research in these forensic pattern disciplines, forensic scientists, academic researchers, and industry partners are transitioning from historically qualitative and subjective practices to approaches that are more quantitative, objective, and data-driven. This can be seen in the trash marks [3], handwriting [4][5][6], footwear [7], firearms and toolmarks [8][9][10][11][12], and latent prints [13,14] disciplines. These recent developments involve automation [9,10], computational algorithms [4,12], and statistical metrics and measurements [14].…”

“…Traditionally, an FDE employs a manual trash mark examination method using appropriate lighting and magnification to record trash marks. However, a semi-automated technique called the electronic Photocopy Examination macro (ePhEx) was recently developed to introduce objectivity into the examination process and increase efficiency [3]. This technique employed a Visual Basic for Applications based macro that communicates between Microsoft Excel™ and Adobe® Photoshop® to detect, locate, record, and image each trash mark.…”

“…Pixels above the selected threshold were assigned to white, and those at or below the threshold were assigned to black. Clusters of co-located dark pixels were referred to as "blobs" to minimize noise [3]. Photoshop tools used within the macro located the black pixels identified the bounds and sizes of each blob, sub-imaged each qualified blob, and recorded relevant information in Excel.…”