India accounts for nearly 40% of low-birth-weight babies in developing countries, and approximately 75% of neonatal deaths in the country occur during the first week of life.
The Challenge
While the Home-Based Newborn Care (HBNC) program under the National Health Mission (NHM) mandates scheduled home visits by frontline health workers (FLWs) to monitor newborn growth, delivering postnatal care comes with real challenges: long hours on foot, difficult terrain, and the physical strain of carrying equipment like weighing scales, infantometers, and measuring tapes.
Our AI Solution
Shishu Maapan is an AI-powered Newborn Anthropometry (NBA) solution that enables FLWs capture newborn measurements using an entry-level smartphone in community settings, for newborns aged 0 to 42 days.
The video-based, easy-to-learn workflow simplifies conventional protocols, supports accurate, digitized newborn records in real time, and helps health workers and systems make timely referrals and corrective actions.
By digitizing the process, Shishu Maapan strengthens proactive care and builds a more responsive and efficient postnatal care system, enabling prompt, accurate measurements at home, minimizing physical contact, reducing reliance on heavy equipment, and safeguarding privacy (videos used for measurement are not stored on the FLW’s device).
Shishu Maapan was awarded Best Paper at the 7th International Workshop on Computer Vision for Physiological Measurement (CVPM) 2024, and Silver in the “Health & Sanitation” category at the Mint Tech4Good Awards, November 2024.
How it Works
01
Video Capture
An FLW captures a short 10–15 seconds video of the newborn on a flat surface with a wooden ruler for measurement reference; the tool works offline for low-connectivity and remote areas.
Video Capture
An FLW captures a short 10–15 seconds video of the newborn on a flat surface with a wooden ruler for measurement reference; the tool works offline for low-connectivity and remote areas.
02
AI Analysis
A Convolutional Neural Network (CNN) extracts and analyzes key anthropometric measurements from the video in real time.
AI Analysis
A Convolutional Neural Network (CNN) extracts and analyzes key anthropometric measurements from the video in real time.
03
Anthropometric Measurements
Using a multi-task learning framework, the solution calculates weight, length, head circumference, mid-upper arm circumference, and chest circumference.
Anthropometric Measurements
Using a multi-task learning framework, the solution calculates weight, length, head circumference, mid-upper arm circumference, and chest circumference.
04
Data Integration
Public health administrators access the measurements through dashboards for real-time monitoring and corrective action.
Data Integration
Public health administrators access the measurements through dashboards for real-time monitoring and corrective action.
Where It's Deployed
Deployed across the Union Territory of Dadra & Nagar Haveli and Daman & Diu (DNH & DD)
1
Union Territory
Attribution and Impact
- Attribution is established through external evaluation assessing implementation and programmatic impact among ASHA workers in DDNH.
- Strong Alignment and Integration With Existing System Delivery Platforms: The solution shows near-universal usage, with 98 percent of surveyed ASHAs (n=82) reporting use of the application in more than 90 percent of their HBNC visits.
- Reduced Operational Burden and Improved Efficiency: 87 percent of users cited ‘clarity of prompts’, 89 percent reported faster measurement, and 88 percent reported ‘ease of mobility’ during field visits.
- Higher Performance: User perceptions indicate significantly higher performance of Shishu Maapan compared to traditional methods across key usability dimensions, with average scores of 8.6 (accuracy), 9.5 (ease of use), 9.2 (efficiency), 9.8 (reliability), and 9.9 (confidence), compared to lower scores for the Salter scale on a 0-10 scale.
- Increased Supervisory Support: 87 percent of users reported receipt of monthly supervisory support, and 99 percent reported increased confidence due to supervisory engagement, including real-time troubleshooting through WhatsApp groups.
Our Other Solutions
Partners
