Application
HaatVisionAI
Smartphone CV pipeline for reference-free 3D food volume estimation at rural market scale.
95%
Classification
14%
Volume error
Mask R-CNNResNet18LSTM6DoF pose
Measures quantity and variety of nutri-dense foods in rural weekly markets. Mask R-CNN (ResNet-18) segments food types; an LSTM IMU network predicts camera pose for marker-free 3D volume estimation.