ITS News

Saturday, April 20, 2024
January 01, 2019 10:01

SIFARS, a Fast and Accurate Face Recognition System

Oleh : itsmis | | Source : -

Prof. Dr. Ir Mauridhi Hery Purnomo MEng (seated) and Dr. Supeno Mardi Susiki Nugroho ST MT discussed the workings of SIFARS

ITS Campus, ITS News – Detecting someone’s face from a CCTV (Closed Circuit Television) camera now no longer needs to be done conventionally, because the research team from Department of Computer Engineering, Institut Teknologi Sepuluh Nopember (ITS) Surabaya created a new breakthrough called SIFARS (Smart ITS Face Recognition System).

The research team consisting of lecturers and students, the team members are Prof. Dr. Ir Mauridhi Hery Purnomo MEng, Dr. Supeno Mardi Susiki Nugroho ST MT, Dr. I Ketut Eddy Purnama ST MT, Dr. Reza Fuad Rachmadi ST MT, Willy Achmad Fauzi and Tahta Dari Timur. SIFARS is claimed to be able to detect the presence of someone faster than the manual observation of a CCTV camera / IP camera.

“For example, in one area there are thousands of CCTVs, this will certainly take a long time to detect someone’s identity,” said Supeno Mardi Susiki Nugroho, a member of the research team.

Prof. Dr. Ir Mauridhi Hery Purnomo MEng (seated) and Dr. Supeno Mardi Susiki Nugroho ST MT with the SIFARS device

Regarding how it works, the man who is familiarly called Uki explained, at first SIFARS worked by take a photo of the person that wanted to be recognized via CCTV. Furthermore, the image data is analyzed by the server and then detected by deep learning technology.

Armed with the principle of deep learning technology, the team processes the flow of frames both in real time and from CCTV footage to recognize someone’s face being sought. “As a result, this system can be used without the need to make new rules, so that it can detect it quickly,” said the lecturer that was born on March 13, 1970.

Dr. Supeno Mardi Susiki Nugroho ST MT (left) and Dr. I Ketut Eddy Purnama ST MT with SIFARS background

Not only superior in terms of speed in detecting faces, Uki also claims that if the output of SIFARS can be accurate up to 99 percent. “Both are seen in full face and half face,” he said.

He added, the breakthrough of SIFARS was also considered by his team to be more compatible. “It’s because SIFARS can be connected with various types of IP cameras,” said the ITS alumnus.

Representing the team, Uki hopes that in the future this SIFARS can be cooperated with the government and the police. “Knowing there’s so many benefits of SIFARS, such as forensic evidence, security monitoring and assisting in tracking someone’s existence,” he assured. (cha/ITS Public Relation Officer/Anjani)

Monitor display of SIFARS detection results

Related News