A scatter search-based technique for pair-wise 3D range image registration in forensic anthropology

Abstract

Different tasks in forensic anthropology require the use of three-dimensional models of forensic objects (skulls, bones, corpses, etc) captured by 3D range scanners. Since a whole object cannot be completely scanned with a single image, multiple scans from different views are needed to supply the information to construct the 3D model. Range image registration methods study the accurate integration of the different views acquired by range scanners, with pair-wise approaches progressively processing every adjacent pair of scanned views until reconstructing the whole 3D model of the object. Our proposal is based on the adaptation of our previous work (Cordon et al, IEEE ) in order to apply the scatter search evolutionary algorithm to pair-wise image registration in forensic anthropology applications. To measure the performance of this adaptation, we design an experimental setup considering some of the most recent and accurate evolutionary techniques for the problem, as well as one skull from our Physical Anthropology Lab. Two additional volumes, commonly used in other pair-wise range IR contributions, have also been considered to complement the comparison of results among the proposals.

Fuzzy-genetic optimization of the parameters of a low cost system for the optical measurement of several dimensions of vehicles

“…In the new system, when the operator has finished selecting the points that are to be matched, geometric transformations will be applied to the images so that the operator's selection can be fine tuned by the computer. The tuning will consist in genetically maximizing the local matching between the set of transformed images in the neighborhood of these points (Santamaría et al 2007). Afterward, the results will be presented to the operator so that he validates the tuned points (or decides to reject the changes).…”

“…This arrangement still allows us to use stereoscopic vision techniques to find the spatial coordinates of the points involved, but prevents the application of classical matching algorithms (Bhat and Nayar 1998;Faugeras et al 1993;Hoff and Ahuja 1989;Scharstein and Szeliski 1998;Tomasi and Manduchi 1998) to pairs of images. There are recent approaches (Santamaría et al 2007) that use metaheuristics and genetic algorithms to find the best correspondences. In our case the images that form a stereo pair of are very different between them (the cameras are spaced about 10 meters apart, see Fig.…”

Replicator Dynamics in the Iterative Process for Accurate Range Image Matching

“…Instead of randomly selecting and combining the candidate solutions, the scatter search (Santamaria et al 2007) adopts for computational efficiency a complete combination of all the possible solutions in a reference set whose size is much smaller than that of the population as in the traditional GA. To combine different solutions, the BLX α method is employed.…”

“…The aforementioned automatic range image matching algorithms can be classified into three main categories: (1) feature extraction and matching (Johnson and Hebert 1999;Huber and Hebert 2003;Brusco et al 2005;Ashbrook et al 1998;Chang et al 2004), (2) iterative possible correspondence establishment and camera motion estimation from these possible correspondences (Besl and McKay 1992;Zhang 1992;Chen and Medioni 1992), and (3) iterative camera motion search and evaluation of the resulting camera motion parameters (Lomonosov et al 2006;Silva 2005;Santamaria et al 2007). All these approaches have their own advantages and disadvantages and could succeed in one situation, but degrade catastrophically in another.…”

Range image registration using hierarchical segmentation and clustering

“…As a result, a large number of automatic algorithms have been developed in the literature. These algorithms can be classified into three main categories with regard to the order between the establishment of possible correspondence and the evaluation of camera motion parameters: (1) feature extraction and matching [5], [4]; (2) an optimal combination of points [3], [2], and (3) camera motion search and its evaluation [10], [13]. Different categories classified here are not exclusive from each other.…”