NCJ Number
195247
Journal
Journal of Forensic Sciences Volume: 47 Issue: 2 Dated: March 2002 Pages: 285-292
Date Published
2002
Length
8 pages
Annotation
This paper provides an overview of the various available image databases and ways of searching these databases on image contents.
Abstract
This research assesses how the various algorithms for correlation and image matching are applicable in forensic science. First, an overview is provided of current research and applications, followed by an application of these methods to several forensic databases, each focusing on a different method of matching images. These databases are tool marks, shoeprints, cartridge cases, and drugs/pills. In the literature, databases are divided into different generations. This research focuses on second generation visual information systems. In these systems, there are different ways of searching in the database. The user can search in the database on features such as texture, shapes, and color distribution. The features can be combined with text strings in the database. With this method, the user can search for a certain group of cases in a forensic database and compare the images with features of the images. Most current research focuses on finding features in images, indexing a database in an efficient manner, and the man-machine interface. Searching in 3D-databases and video-databases is getting more attention in the literature. New developments in this field will not only include faster systems, but also standardization of the implementation of new algorithms will occur. In 2001 the MPEG-7 standard will be a standard for indexing images and video. This might have an impact on the way searches are conducted for video and audio on the Internet. A sketch generated by the user will be compared with a database of images. Whether this framework will actually be used very much will depend on the market. In future databases, more powerful methods that require parallel processing can be used. 3 figures and 29 references