Calls of “Roll number 1, 2, 3…” may soon become a thing of the past in Telangana schools as the government floats a tender to replace roll-calling with a facial recognition system for attendance.
As per the tender, the government is working to create a facial recognition attendance management solution to automate the identification and verification of students and teachers. The system will be fed ‘facial templates’ of the persons being scanned at the time of user enrolment.
Why it matters: Although there is no law in India for data protection and user privacy, the Digital Personal Data Protection Bill, 2022, states that the collection and processing of children’s data require parental consent. However, in the current document, there is no mention of the parent’s, student’s or teacher’s consent. Facial data can be considered sensitive data especially if collected over an entire academic career. While the tender asks the bidder to “adhere to all data privacy guidelines,” what measures will the entity take if there are no set safeguards to protect children’s data?
Additionally, this tender should also be looked at alongside the proposed plan for an ‘Education Ecosystem Registry’ by India’s Education Ministry that seeks a unified system to track learner outcomes and inform policy-making processes in the education sector. If this system is connected to the ‘unified’ system, there are further privacy and data risks to consider as highlighted in a recent consultation paper.
STAY ON TOP OF TECH POLICY: Our daily newsletter with top stories from MediaNama and around the world, delivered to your inbox before 9 AM. Click here to sign up today!
Facial recognition for group attendance
The attendance system will work at two levels, with an option for group attendance by the teacher or self-attendance. In case of group attendance, a set of individuals will be photographed in one picture.
One person nominated by the department will have login access for a class or multiple classes. As per the tender, the solution should be able to support a group of at least 20 individuals.
“Attendance should be always taken in group where images or video should not be stored in the devices. Facial recognition solution should find the person identity based on facial points comparison,” said the tender.
The person in charge will have a unique end user access level feature to link/unlink related enrolled images. Further, the attendance will be able to take false rejection cases in the same session and update the attendance reporting accordingly. The system would be in charge of the attendance of around 20 lakh individuals.
Facial recognition for self-attendance: At the individual level, students can take their own attendance by clicking their own selfies. A designated person will oversee the enrolment process or can take the student’s photo for them.
However, as per the tender, the system will also be built to block “duplicate enrolments” of the same candidate while allowing twins. The tender doesn’t mention how the system will do so. As with group attendance, the system will not store the daily selfies taken by the students.
Facial recognition for teachers: The same system will be used to mark the attendance of teachers as well. Teachers will use their mobile phones to take selfies and run them through the facial recognition system application installed on their mobile.
“Authorized person/ Individual points the camera within the app towards the Individual to whom need to take the attendance and scans the individual face. The app recognizes faces using cognitive AI functions and marks the attendance for the individual,” said the tender. Unlike the system features set for scanning students, there is no provision for group facial recognition of teachers.
How will the system work practically?
As per the tender, the user-assigned class teachers will mark attendance. Attendance will be taken “one or multiple times in a day” based on “the decision of user department”. Still, the attendance system should go as follows:
“I. Faculty [class teacher or assigned person] enters the class and starts the App.
- Faculty enters the classroom from the schedule for which the attendance needs to be taken in physical location.
III. Faculty authenticate and register their own attendance using the front camera.
- Faculty points the camera within the app towards the class and scans the individual faces.
- The app recognizes faces and marks the attendance of the user…
VIII. The solution should be able to determine multiple Headcount/Faces from a single frame of class attendance during attendance capturing time.”
Individual attendance must be completed within 10 seconds while group attendance should be completed within 120 seconds per frame, said the tender.
Attendance system to also make an absentee list: Based on the faces scanned and an enrolment list prepared beforehand, the system will also show all the absentees of the class along with their images and count of absentees. It will also create an absentee list.
Facial recognition system to raise alerts to teachers: The tender also called for the attendance system to alert the relevant authority and generate reports if a student’s attendance drops below the required mark. Similarly, the system must alert the faculty if there is continued absence of a student or if the student arrives late to the class. This will be done via SMS/WhatsApp notification.
Attendance data created based on facial recognition system: The new system will be integrated with the applications shared by other user departments to get master data regarding attendance. These data sets will be granular enough to contain detailed reports at the Institute/ Mandal/ District/ State/ Department level based on the requirements of the user department. Similarly, the system will create daily, week-wise, month-wise and summary reports. In case of internet disconnection, the data shall be available offline and shall be synced to the application later.
System to store data for up to five days: The facial recognition app will store the attendance data for up to a week (five days) before sending it for mandatory data synchronization with “the central server”.
“Under normal conditions, the data will synchronize with the central server whenever there is connectivity,” said the tender.
This post is released under a CC-BY-SA 4.0 license. Please feel free to republish on your site, with attribution and a link. Adaptation and rewriting, though allowed, should be true to the original.
- Mumbai To Get Hyderabad-Like FRS-Based Surveillance To ‘Curb Street Crime’
- Deep Dive: A Plea Against “Coerced” FRS Attendance, And The Harms Of Workplace Surveillance
- Indian Government’s DIKSHA App Under Scrutiny For Alleged Breach Of Children’s Data
- How The Data Protection Bill Restricts Children’s Access To The Internet #NAMA