Tirupati temple calls EoI for AI based facial recognition, pilgrim headcount solutions

Tirumala Tirupati Devasthanams ( TTD) is set to implement solutions for AI based facial recognition, head count of pilgrims and live detection of faces. The temple authorities have issued an Expression of interest ( EoI) from vendors to present products and solutions before the TTD committee on July 31.

TTD plans to implement Facial Recognition based counter service for Slotted Sarva Darshan, Sarva Darshan & Current Booking Accommodation to identify the pilgrim uniquely and to extend the service to the same person. As part of FR process , live photos of pilgrims will be captured using web camera from the TTD software application. This will be checked with pilgrims' faces recorded over one-month to identify the duplicates of about 10 lakh faces.

In case duplicate is identified, the application is rejected to provide the connected booking service and in case of no duplicate being identified, the pilgrims will be provided with darshan/accommodation token. There are about 40+ counters running parallelly in SSD Counters, 10+ counters in Sarva Darshanam and 10+ counters in Current Booking Accommodation. Facial Recognition Solution will be Artificial Intelligence enabled and will capture photo of pilgrims with liveliness validation.

AI enabled Camera system will be used for live detection of faces. The pilgrim photos of various booking services like Online Special Entry, Seva, Seva DIP and Offline bookings like Slotted Sarva Darshan, Sarva Darshan are being captured through various processes which is about 1 lakh records per day. The live feed or images of group of people will be given as input and the camera system will identify a particular pilgrim whose image is already available in the database. This will help assess unauthorized entries in the system, time taken for darshan category wise on a live basis etc, according to the tender document.

TTD also plans an AI based head counting solution at Srivari temple which counts heads with highest accuracy so as to assess the time for darshan, darshans actually performed compared to planned etc. This will help identify the exact number of pilgrims who had darshan during a particular time period. The document notes that the queue lines and movement of people is very complex because of which TTD tried many solutions but could not get desired head count accuracy.

TTD allows pilgrims to wait at Vaikunta Queue Complex which has about 30 compartments (wide open hall) and the capacity of each compartment is about 400 to 450. TTD will introduce a solution to capture the live situation of the all the compartments individually and other aspects like how many of them have been filled with pilgrims and how may are vacant. FR technology was implemented on a trial basis in 2023.