At the Swedish National Forensic Centre [Nationellt Forensiskt Centrum] (NFC) 3D-models are being used to document crime scenes, and visualizations based on 3D-models are being used in court. In addition to this, 3D-models of crime scenes allows for new methods to be developed and used. One of these are position estimations in 3D-models based on 2D-images. The software program P3KA has been developed at NFC for the purpose of conducting assisted position estimations with uncertainty analyses in 3D-models. The objective of the current study was to develop guidelines for the design and presentation of the uncertainty analyses produced by P3KA. The research questions that the current study aimed to answer were as follows. - How should uncertainty analyses for position estimations in crime scene visualizations be designed and presented? - What guidelines can support the process of designing and presenting uncertainty analyses for position estimations in crime scene visualizations? The concept of uncertainties is multifaceted and can be defined to include both statistical uncertainties and the uncertainty of scientific judgement. Uncertainties are unavoidable, and the field of forensic science is no exception. Previous studies have investigated the use of heat maps as a strategy for representing uncertainty. This strategy was applied to represent the uncertainty analyses produced by P3KA in an effort to create intuitive and comprehensible representation of uncertainty. In addition to studies regarding the representation of uncertainty, the field of information visualization provided a theoretical foundation on which the first version of guidelines could be developed. The guidelines were then evaluated through an iterative process consisting of three focus groups with members of the public, journalists, and communicators, as well as eight interviews with actors working within the justice system. All these groups have different relationships to crime scene visualizations, and could therefore contribute with knowledge about the domain, and different perspectives regarding the representation of uncertainty analyses for position estimations in crime scene visualizations. This was important to ensure that the representation was interpreted as intended. The guidelines were adjusted between each iteration, and finally one last version of the guidelines was established. Thus, answering the second of two research questions. The results from the current study furthermore showed that uncertainty analyses for position estimations in crime scene visualizations should be designed as single-colored heat maps that use a change of hue and boarders to distinguish levels of uncertainty within the representation. There should also be a clear marking that indicates the estimated position. When an uncertainty analysis is presented, it should be accompanied by the purpose of the position estimation; the results from the position estimation; the conclusions that can be drawn based on the position estimation; and the context for the position estimation. A crime scene visualization containing a position estimation should also be accompanied by documentation of the procedure. Additional views of the visualization or scales can be added to support the recipients understanding for spatial relationship and distances at the scene of the crime. Finally, the elements that are included in the crime scene visualization containing a position estimation and an uncertainty should be explained, through a legend for example. By creating guidelines for the design and presentation for representations of uncertainty analyses for position estimations, a standardized and transparent process can be achieved. Thus, supporting the legal security of the evidence that position estimations and crime scene visualizations constitute. The guidelines also open for a discussion regarding bias in the field of forensic science.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:liu-185732 |
| Date | January 2022 |
| Creators | Iversen, Katarina |
| Publisher | Linköpings universitet, Institutionen för datavetenskap |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0023 seconds