Design And Development of Interactive Mirror for Aware Home.
Chidambaram Sethukkarasi, Vijayadharan SuseelaKumari HariKrishnan, Raja Pitchiah
National Ubiquitous Computing Research Centre
Centre for Development of Advanced Computing
Chennai, India
{ctsethu, harikrishnans, rpitchiah}@cdac.in
Literature Survey
The Interactive Mirror comprises of a dielectric coated mirror mounted over a LCD Display, a camera for capturing the user’s image, load sensors for measuring user’s weight, Radio-frequency identification (RFID) reader and RFID tags for identifying the garment worn by the user.
The contributions,
1) Conceptualization,design and development of interactive mirror.
2) 2) facial feature extraction for emotion recognition.
3) 3) Face recognition and emotion recognition algorithm training.
4) 4) Tests results of face recognition and emotion recognition algorithm in the deployed environment.
Related works
Use speech processing techniques.The mirror interacts with the user through verbal commands, listen to user’s question and responds them.
A mirror makes use of behavioral data in order to provide its user with continuous visual feedback on their behavior in a natural manner.
The system captures the image of the surrounding and display on a LCD screen.
The system uses robust multiple face detector and tracker based on active infrared illumination, and developed a face model to generate realistic graphics output,
Engineering of mirror
Dielectric coated mirror of thickness 1/4 inch. Entire panel act as mirror and display. In the top of the mirror web camera having VGA resolution and frame rate of 30fps. Weight measuring platform is formed with 4 load sensors from loadstar. The antenna is inside the wooden frame.
Face recognition
First step of face recognition is face detection. Face detection is finding faces in an image. Detect using Viola and Jones algorithm. It can recognize image face or non-face.
Second step is recognize whose face is. Eigen face recognition method is use here. The face images of the users are captured and stored in a face database. Algorithms trained with these images and calculate the distance between input image.
Emotion recognition
Facial expressions can be recognized from the facial features.
Features of face that change by changing the emotions. 1) Mouth. 2) Eye.
This use K-NN classification algorithm.
Event reminder
User has to enter the details which they want to be reminded using GUI. When user is using the mirror, the system check for reminders display it.