Height Estimation of Children under Five Years using Depth Images

Area of Science:

• Pediatric Health
• Computer Vision
• Medical Imaging

Background:

• Malnutrition is a leading cause of mortality in children under 5 globally.
• Accurate anthropometric measurements (weight, height, MUAC) are vital for malnutrition detection.
• Resource limitations in the Global South hinder accurate anthropometric data collection.

Purpose of the Study:

• To develop a CNN-based method for estimating height in young children using smartphone depth images.
• To provide an accessible and accurate tool for detecting stunting in resource-limited settings.

Main Methods:

• A Convolutional Neural Network (CNN) model was trained on 87,131 depth images.
• Depth images were collected using a smartphone, capturing standing children under 5.
• Model performance was evaluated on 57,064 test images.

Main Results:

• The CNN model achieved a mean absolute error of 1.64% in height estimation.
• Height was accurately estimated within the acceptable 1.4 cm range for 70.3% of test images.
• The model demonstrates potential for accurate stunting detection.

Conclusions:

• Smartphone-based depth imaging with CNNs offers a viable solution for accurate height estimation in young children.
• This technology can improve malnutrition detection, particularly stunting, in resource-constrained environments.
• The proposed method addresses a critical gap in global child health monitoring.