Computer vision methods applicable to forensic science

Workshop of Theses and Dissertations, XXVI Conference on Graphics, Patterns and Images (WTD/SIBGRAPI), Arequipa, Peru, 2013


Crimes in our society, increasing in volume and sophistication, have determined the need for knowledge and use of scientific methods to their prevention and investigation. This work presents three Computer Vision methods that can be applied in forensic investigations. In the first method, a new vanishing point detector facilitates the process of making measurements in a single 2D image, and is used to estimate the height of a person in an image, important measurement to corroborate pieces of evidence. In the second method, multiview stereo techniques are used to obtain a three-dimensional model from photographs taken from a footwear impression, evidence commonly found in crime scenes. In the third method, the need for reconstructing shredded images is explored. A photograph can be shredded in order to hide information and it is up to the field of Computer Forensics its automatic reconstruction. Experimental results are provided, showing the effectiveness of the proposed methods when compared to relevant work in the literature.


    authors      = “Fernanda A. Andal{\‘o} and Siome Goldenstein”,
    title        = “Computer vision methods applicable to forensic science”,
    booktitle    = “Workshop of Theses and Dissertations, {XXVI} Conference on Graphics, Patterns and Images ({WTD/SIBGRAPI})”,
    year         = 2013,
    address      = “Arequipa, Peru”,