Mobile AppMyStressD
MyStressD is a stress-detection mobile app built with Flutter & TensorFlow Lite. Upload a photo or image, and the on-device model analyzes your stress level—no data ever leaves your phone.
View JournalFeatures
On-device Stress Detection
Runs inference fully on-device using TensorFlow Lite—your data never leaves your phone.
Image-based Analysis
Upload a photo or image; the app detects stress cues and returns a stress level score.
Flutter UI
Built in Flutter for a smooth, responsive experience on both Android and iOS devices.
CNN Architectures
DenseNet201
A densely connected convolutional network that mitigates the vanishing-gradient problem, reusing features across layers.
MobileNetV2
A lightweight, efficient model optimized for mobile inference, using inverted residuals and linear bottlenecks.
ResNet50
A 50-layer deep residual network that uses skip connections to enable training of very deep models.
Tech Stack
Flutter
Dart
TensorFlow Lite