wordpress blog stats
Connect with us

Hi, what are you looking for?

What happened at NCRB’s meeting with vendors on Automatic Facial Recognition System

Image representing facial recogntition

Vendors bidding for the NCRB’s (National Crime Records Bureau) centralised automatic facial recognition system (AFRS) raised queries on its future integration with other systems, and asked how it would deal with plastic surgery cases in a pre-bid held by the NCRB on July 25. Internet Freedom Foundation obtained the minutes of this meeting via RTI.

About 80 vendors attended the meeting chaired by NCRB’s deputy director, A. Mohan Krishna. The bureau’s joint assistant director Pawan Bharadwaj was also present.

On June 28, NCRB had called for bids for implementation of a centralised AFRS. Its request for proposal stated that this was being done to modernise the police, gather information, identify criminals, among other things. AFRS would also allow information dissemination among various police organisations and units, said the RFP.

Here’s what happened in the meeting, according to the RTI:

1. Some queries were raised regarding future integration of AFRS with other systems. According to the RFP, AFRS will be integrated with ICJS, CCTNS, Immigration Visa Foreigner Registration Tracking (IVFRT), advanced State police Integration (Police IT Karnatak, Enterprise e-Cops Telangana, G-Cops Goa, Cyprus Tamil Naidu, e- Guj Gujarat). It also needs to be integrated with other biometric solutions deployed at police department system like Automatic Fingerprint identification system (AFIS).

2. How will AFRS identify plastic surgery cases? The vendors’ AFRS systems should be able to broadly match a suspect/criminal photograph and should be able to use images of varying quality levels such as those with: varied lighting conditions, small image sizes (300 x 300 pixels), low Jpeg image quality, plastic surgery, aged images, bearded face images, makeup images, slanted Face, the RFP states.

3. Number of queries raised about camera/concurrent users, Data Centre and Disaster Recovery sites. The RFP said that there is a need to ensure that a Disaster Recovery site is available for testing the application when Data Centre is brought down. Also, the source database for the AFRS should have a capability of handling 2500 concurrent users.

4. Queries about pre-qualification criterion such as past experience and size of assignments in executing similar assignments. According to the RFP, bidders need to send a brief write-up to NCRB on their past experience with similar assignments, highlighting their hardware, software, AFRS Solution, etc.

  • Here’s how they’d be scored according to the RFP: An assignment whose database size is more than 5 million will be awarded 10 marks and 8 marks is proposed for 4-5 million. Similarly, 3-4 million will be getting 6 marks, 2-3 million will be getting 4 marks and 1-2 million will be getting 2 marks. No marks will be allotted for less than 1 million

5. Most participants requested to extend the last date for submission of the bid documents. According to NCRB’s website, the last day to submit bids was initially September 13. However, it has now been extended until October 11.

6. Most participants requested to allow a consortium of three companies to bid. According to NCRB’s 172-page RFP, a joint bid of up to two companies can be made. Also, the joint-bidding should have come into force before the last date of submission of tender.

7. Queries were raised about NIST (National Institute for Standards and Technology)/FRVT (Face Recognition Vendor Test) standards. NIST is an agency of the United States Department of Commerce. FRVT is a standard set by NIST which measures the accuracy and speed of one-to-many face recognition identification algorithms. In the RFP, NCRB said that bidders should “desirably” have participated NIST FRVT evaluation programme 2016. Bidders need to provide NIST documentation showing its scores.

8. Many participants requested to reduce the turnover criteria. Bidders/joint venture companies should have an annual turnover of at least Rs 100 crore in each of the last 3 financial years, according to the RFP. In case of a joint bid, each of the partner should have an annual turnover of at least Rs 100 crore in each of the last 3 financial years.

9. Some participants requested EMD (Earnest Money Deposit) exemption for NSIC (National Small Industries Corporation)/MSMEs. Bidders shall submit, along with their Bids, an EMD amount Rs 40 lakh. EMD is basically a bid security, per the RFP. NCRB, will return the EMD of unsuccessful bidders.

The RTI response says that NCRB will examine all the queries that were raised by the vendors at the public meeting. It will publish answers on its website for them to view. NCRB had requested them to send queries by July 29.

According to a report by The Quint, the issue of algorithms was also raised at the meeting. A vendor submitted that many companies who have a good algorithm for this system were not able to participate in the bidding because of the experience criterion. This detail was missing from the RTI.

Read our detailed coverage on the Automated Facial Recognition System here.

You May Also Like


Following the raging debate around the latest updates to WhatsApp’s privacy policy, RSS ideologue K.N. Govindacharya urged the government to tax data transfers between...


Mobile numbers and WhatsApp chats of people using WhatsApp web were indexed on Google search results yet again, a security researcher claimed. This came...


The Ministry of Electronics and Information Technology wrote a letter to WhatsApp CEO Will Cathcart calling on the Facebook-owned messenger app to withdraw proposed...


Senior journalist and news anchor Nidhi Razdan was all set to start teaching at Harvard University this year. But it turns out she appears...

MediaNama is the premier source of information and analysis on Technology Policy in India. More about MediaNama, and contact information, here.

© 2008-2018 Mixed Bag Media Pvt. Ltd. Developed By PixelVJ

Subscribe to Daily Newsletter

    © 2008-2018 Mixed Bag Media Pvt. Ltd. Developed By PixelVJ