Superimposition

Abstract: In software product line engineering, feature composition generates software tailored to specific requirements from a common set of artifacts. Superimposition is a technique to merge code pieces belonging to different features. The advent of model-driven development raises the question of how to support the variability of software product lines in modeling techniques. We propose to use superimposition as a model composition technique in order to support variability. We analyze the feasibility of superimposition for model composition, offer corresponding tool support, and discuss our experiences with three case studies (including an industrial case study).

Abstract: This paper proposes an image signature robust to caption superimposition for video sequence identification. A new image signature which is a set of local features is developed for a high-speed frame-by-frame matching of video sequences. The signature of a frame is obtained by partitioning the image into blocks and extracting the local feature representing the dominant type of edge direction from each block. The similarity between the signatures is calculated by comparing the edge types of the corresponding blocks, and counting the number of the blocks having the same edge type. A weighting scheme based on the probability of caption superimposition for each block can be applied to the similarity calculation to improve the matching performance. The experimental results of the video sequence identification show that the proposed signature achieves precision of 99.65% and recall of 99.45%, improving both the precision and the recall by more than 30% compared with the conventional signature.

Abstract: A facial image was reconstructed from the skull, part of a complete skeleton found in woodland, of a male person who had hanged himself from a tree. In addition, video superimposition was carried out with antemortem photographs of a person suspected of being the victim, and a good match was obtained. In a further case, a cheaper video-transparency superimposition was carried out, with identity later being confirmed on the basis of dental records. The techniques and the problems encountered are discussed. According to our experience, 3D computer reconstruction and video superimposition have a useful role in the process of identification, particularly in the early stages of an investigation and when other more definitive methods may not be available.

Abstract: The accuracy of video superimposition methods for identifying unknown human skulls was examined. Three identified human skulls were each compared to 97 lateral view and 98 frontal view “mug shot” photographs using two television cameras, an electronic signal mixer, and a video monitor. The skulls were not from individuals represented by the photographs. All comparisons were done without using anterior dentition. The results found that 9.6% of the lateral view and 8.5% of the frontal view superimpositions were classified as a consistent fit based on the criteria that were identified. The incidence of false matches was reduced to 0.6% of the sample when a frontal view and lateral view photograph of the same individual were both compared to one skull. It was concluded that without anterior dentition, skull/photograph superimposition is reliable when two or more photographs, clearly depicting the facial features from different angles, are used in the comparison.