KinderAct is an AI-powered web application developed on the Google Cloud Platform (GCP) that specializes in real-time facial expression recognition using Vertex AI and App Engine technologies. The project focuses on enhancing user experience and interaction by accurately detecting and interpreting facial expressions.
- Real-time facial expression recognition from camera feed.
- Utilization of Vertex AI and App Engine for seamless deployment and scalability.
- Custom Python scraper employed for scraping and labeling 150+ facial image datasets.
- Achieved over 90% accuracy rate in labeling emotions.
- Trained a computer vision deep neural network model with over 82% validation accuracy in detecting 3 primary facial expressions: happy, sad, and angry.
- Google Cloud Platform (GCP)
- Vertex AI
- App Engine
- Python for scraping datasets
The project utilized a custom Python scraper to collect and label a diverse dataset comprising over 150 facial images. This dataset was meticulously labeled to ensure high accuracy in emotion recognition.
The computer vision deep neural network model was trained using the collected dataset. The model achieved an impressive validation accuracy of over 82% in accurately detecting happy, sad, and angry facial expressions.
To utilize KinderAct, simply access the web application through the provided link. Allow access to your camera, and the application will seamlessly recognize and interpret your facial expressions in real-time.
- Open KinderAct App
- Click "Mulai"
- Allow your camera access on your browser
- Enjoy!!
Notes: The image on the left shows an example of an expression that can be followed
Demo Project (https://youtu.be/irVxBsLXkRU)
