EmotiCalm is a facial recognition-based application designed to detect stress levels and provide personalized suggestions for stress management. The project integrates cloud computing, mobile development, and machine learning to deliver a seamless experience for users.
EmotiCalm aims to leverage advanced technology to assess stress levels from facial images and provide actionable suggestions. This project consists of three primary components:
- Cloud Computing (API): Manages stress detection models and serves predictions.
- Mobile Application: Allows users to capture images and receive results and suggestions.
- Machine Learning: A TensorFlow-based model to analyze stress levels from facial features.
Below is the system architecture for the EmotiCalm project:
- Google Cloud Storage: Stores machine learning models and user-uploaded images.
- Cloud Run: Hosts the API for stress prediction.
- Firebase: For Handle user authentication.
- Mobile Frontend: A user-friendly interface to interact with the system.
- Machine Learning Model: A TensorFlow-based neural network for facial recognition and stress detection.
The mobile application design focuses on user accessibility and clarity. The interface includes:
- Image Capture and Upload: Allows users to upload facial images.
- Stress Analysis Display: Shows detected stress level and corresponding suggestions.
- History Tab: Keeps track of past analyses and suggestions.
Below the screnshoot of our application
The stress detection model uses a convolutional neural network (CNN) trained on annotated facial image datasets. The design involves:
- Data Preprocessing: Facial image normalization and augmentation.
- Model Architecture: Multi-layered CNN for feature extraction and classification.
- Label Mapping:
- 0: No Stress
- 1: Weak Stress
- 2: Medium Stress
- 3: Strong Stress
To view specific implementations, refer to the following branches:
- Cloud Computing (API):
branch-cc - Mobile Development:
branch-md - Machine Learning:
branch-ml