Intelligent Indexing of Crime Scene Photographs

first find a crime scene, the objects and subjects as well as their spatial relations and conditions, all are crucial for collecting evidence and for drawing conclusions during crime investigation. Although investigating a crime is a time-consuming process, the crime scene cannot be preserved for long: life must take again its normal course, objects must be removed, the space must be cleaned and cleared. As these inevitable changes occur, the risk of contaminating the scene and destroying possible evidence grows. Therefore, scene-of-crime officers take a series of photographs as soon as they arrive at a crime scene, and they create a photo album for each case. Each photo album’s first page is an index consisting of a caption for each photograph or set of photographs numbered in sequence. This visual documentation, the official reports of the scene-of-crime officer’s actions, and the evidence collected from the scene are the crime investigator’s main sources of information. However, to retrieve information from past cases or to uncover possible similarities and patterns among cases, current practices rely largely on either the investigator’s memory or his or her availability to go through piles of case files and photo albums. During the last decades, law enforcement agencies have made many attempts to bring information technology to bear on crime investigation. In Britain, the British Police Information Technology Organisation (PITO) and various software companies have developed management systems to facilitate the administrative aspects of crime investigation.1 These systems, currently under pilot testing in various police forces, allow monitoring and control of document flow throughout the investigation, visualization of the sequence of events, automatic population of official reports with verbal information provided by the officers, evidence tracking along the whole custody chain, and task and exhibit management. These systems can also store photographs and other caserelated information in a central database and allow their retrieval through case-related keywords. Users trace photographs either through their unique case number or through information specific to the case, such as the scene-of-crime officer’s name, the type of offense, the date, and the crime scene location. Indexing and retrieving photographs and other case documentation this way will clearly change current practices and facilitate crime investigation. However, intelligent support for this task could take investigation itself—rather than its administration—a step further. Intelligent, automatic indexing and retrieval of crime scene photographs is one of the main functions of SOCIS, our research prototype developed within the Scene of Crime Information System project.