Home / Technology / Two Class 10 students build ₹23,000 AI system to detect child-safety incidents at Hyderabad day-care centres

Two Class 10 students build ₹23,000 AI system to detect child-safety incidents at Hyderabad day-care centres


Two Class 10 students build ₹23,000 AI system to detect child-safety incidents at Hyderabad day-care centres
Meet Sajag AI: Students from Hyderabad and Noida develop AI-powered child safety system using CCTV cameras. (AI Image)

Two Class 10 students from Hyderabad and Noida have developed an AI-powered system designed to detect potentially unsafe incidents involving children at day-care centres. Named Sajag AI, the computer-vision technology analyses existing CCTV footage to identify possible incidents and automatically alerts centre coordinators and monitoring teams. The system was developed by Shloak Bihani of CHIREC International School, Hyderabad, and Aryaman Taparia of Step By Step School, Noida. The prototype reportedly cost around ₹23,000 to develop.Sajag AI has been introduced at COWE India’s Aveksha day-care initiative in Hyderabad, which is being implemented in collaboration with the United Nations Development Programme (UNDP). Media reports state that the technology was demonstrated at the launch of the partnership on November 13. The initiative currently operates two pilot centres in Pashamylaram and Suraram, with plans to expand the AI-based monitoring system to 18 centres.

How does Sajag AI detect child-safety incidents?

Sajag AI uses computer vision to analyse footage captured by CCTV cameras already installed at day-care centres. Rather than requiring centres to replace their existing surveillance infrastructure, the system is designed to work with available camera feeds.When the software identifies an incident that may involve unsafe or potentially abusive behaviour, it sends an alert to the centre coordinator and monitoring team through SMS and email. The notification includes incident details and a link to a snapshot, allowing the team to review the footage and respond.The system is designed to examine multiple consecutive detections before triggering an alert. This approach is intended to reduce false alarms that may arise from isolated movements or routine interactions between children and caregivers.

Built using Raspberry Pi 5, prototype costs around ₹23,000

The students developed the prototype using a Raspberry Pi 5, a compact single-board computer used for computing and hardware-based projects.According to the project details, the AI model was trained using more than 11,000 labelled images to identify potentially unsafe situations. The students reported an incident-identification rate of 97.3% in a separate held-out test.However, live testing also produced false alerts, indicating that the system still requires monitoring and refinement. The reported test performance should not be interpreted as a guarantee that every incident will be detected or that every alert will indicate an actual safety violation.

Sajag AI introduced at COWE India’s Aveksha day-care centres

The technology was introduced as part of COWE India’s collaboration with UNDP to strengthen workplace-based childcare facilities in Telangana.According to media reports, two Aveksha centres, located in Pashamylaram and Suraram, will operate as pilot facilities under the partnership. Each centre is expected to provide three nutritious meals a day along with early learning education.The November 13 launch was attended by Shruti Ojha, Special Secretary and Director of Women and Child Welfare and Social Welfare; Jeya Chandran, Regional Head for Southern India at UNDP; and representatives of COWE India.The Sajag AI system is intended to provide an additional layer of monitoring at these facilities by helping staff identify incidents that may require immediate attention.

How the AI system could support day-care monitoring

CCTV cameras are commonly used in childcare facilities, but continuously watching multiple camera feeds can be difficult for staff responsible for several children and daily operations.Sajag AI is designed to analyse the footage automatically and notify designated personnel when its system identifies a potentially unsafe incident. The alerts are intended to help monitoring teams review situations without having to watch every camera feed continuously.The students’ prototype has been introduced as part of the Aveksha initiative, with a proposed expansion across 18 centres in Hyderabad. Its reported cost of approximately ₹23,000 also makes it a relatively low-cost technology prototype compared with more extensive surveillance infrastructure.



Source link

Tagged:

Leave a Reply

Your email address will not be published. Required fields are marked *