Breaking News

Gujarat helped trace Rajiv Gandhi killers

AHMEDABAD: Twenty-one years ago on this day, former prime minister Rajiv Gandhi was assassinated in Sriperumbudur in Tamil Naduby the first human bomb in the world - anLTTE woman militant called Thenmozhi Rajaratnam alias Dhanu. Architects of the assassination plot Sivarasan, Nalini and Murugan remained elusive from the special investigation team. 

However, it was the portrait-building software developed in the Physical Research Laboratory(PRL)-Ahmedabad and later upgraded at Dharamsinh Desai University (DDU) that cracked the case in a month by drawing near-similar features of the criminals from witness statements. The software was called the Facial Analysis and Criminal Identification System (FACIS Project III) 

Professor H S Mazumdar, who heads the research wing at DDU, was working with PRL's electronics department on the day of Rajiv's assassination. "Initial attempts to put together faces of suspects did not head anywhere. So we developed an indigenous software," says Mazumdar. 

"A large number of photographs of criminals were scanned. From these, we digitally extracted facial features which could be quickly selected and pasted in any desired location to assemble a portrait. It was the first of its kind in the country at the time." 

Software that helped crack Rajiv Gandhi assassination upgraded 

Imagine just dragging a button across a 1980 photograph of terrorist Dawood Ibrahim on your computer screen and being able to witness how he would look when he turns 65 years. Drag the button backwards to know how Ibrahim looked when he was 30 years. 

This is no animated junk that you see on advertised web pages but is an upgraded version of the portrait-building software - the very same that cracked the Rajiv Gandhi assassination case in 1991. The upgraded version of the software developed by the Dharamsinh Desai University's R&D division students will soon be part of NCRB and police department's arsenal across the country to identify criminals who are absconding over a long period of time. 

"The software works as a profiler of criminals. From a previous file photograph of the criminal, the software picks those important points on the face that usually change in texture and tone during the course of time. The button is dragged across the timeline in the software to witness the face of the criminal at different ages. The software has a database of 50,000 digitised images of criminals to choose the facial features from," says head of the R&D wing of DDU university M H Mazumdar

"The classification of these images is the soul of software. As the witnesses use to describe the culprit like 'old', 'weak', 'drooping face', the software immediately brings forth related images and can even morph a set of images on the spot to reach a probable face. The timeline helps the investigator see how criminal would look in different years of his age," says Mazumdar. Students have also armed the software with features that can also have commercial applications. If installed in a bank the software can scan a face from a CCTV footage and immediately tell whether the person is an outsider or a bank customer from the existing database of the bank.

No comments