| dc.description.abstract | Crimes are committed in society all around the world, and some crimes also disturb the peace and order of society as well. It is important to identify the suspects quickly and catch the criminals and then conduct trials to put them behind bars as a punishment when necessary, especially when the crime is severe. However, accurate and real-time person identification still remains to be a problematic issue, even if digital images of the crime scene are available. With the available information (CCTV images, stored images of suspect etc.) in this regard, we have learned that it is possible to make an accurate identification of a person if object recognition (tattoos, necklace etc.) is done along with face recognition. This holistic approach would be more reliable in multiple aspects of the same person or crime scene are verified. In our work, we have designed and developed a computer-based system, called FODET for determining person or objects involved in crime, with new ideas that make this tool very useful in practice. FODET is the user-friendly tool and designed to meet above challenges while helping in identifying suspects and objects involved in crime with the high level of accuracy. In addition to face recognition, detailed object recognition along with image differentiation, boundary based recognition is proposed and implemented in our work. We have conducted tests to show that the results are accurate, and the system can also be configured to suit the needs of a criminal investigation department. We believe that our work has a lot of practical application and it could be adopted in the future by state police or crime investigation departments. | en_US |
