1) The document presents a real-time static hand gesture recognition system for the Devanagari number system using two feature extraction techniques: Discrete Cosine Transform (DCT) and Edge Oriented Histogram (EOH).
2) The system captures an image using a webcam, performs pre-processing, extracts the region of interest, then extracts features using DCT or EOH before matching against a training database to recognize the gesture.
3) An experiment tested 20 images and found DCT achieved a higher recognition accuracy of 18 gestures compared to 15 for EOH.