Retinal Recognition

This project was funded by National ICT R&D Fund as a grass root UG level project.

Retinal recognition based security systems are very rare due to retina acquisition problem but still it provides the most reliable and stable mean of biometric identification. Automated retinal recognition system is a four stage personal identification system using vascular pattern of human retina. In first step, it acquires and preprocesses the colored retinal image. Then blood vessels are enhanced and extracted using 2-D wavelet and adaptive thresholding respectively. In third stage, it performs feature extraction and filtration followed by vascular pattern matching in forth step.

The system offers an improved visibility to cardiac specialist in terms of coronary anatomy and myocardial perfusion, and a single platform in reaching the decision with greater certainty. Two-in-one ‘Hybrid’ CT-SPECT systems of different companies are available in the market abroad but they cost upto Rs 8 to 10 Crore. The system designed attempts to unify the results of two separately existing non-invasive diagnostic modalities, as one, in order to get same or comparable information as that hybrid machine generates, without any extra financial/human resource burden to the institutions.