Hand Gesture Recognition and Interface via a Depth Imaging Sensor for Smart Home Appliances

This paper presents a novel hand gesture interface system via a depth imaging sensor for appliances control in smart home environments. To control appliances with hand gestures, we recognize the hand parts in a depth hand silhouette and generate control commands. In our methodologies, we first create a database (DB) of synthetic hand depth silhouettes and their corresponding hand parts-labelled maps and then train a random forests (RFs) classifier with the DB. Via the trained RFs, we recognize the hand parts in a depth silhouette. Then based on the information of the recognized hand parts, control commands are finally generated according to our implemented control interface. By testing our system on real hand gestures, we have obtained an average recognition rate of 98.50% from five different subjects. With the presented interface system, users can control smart home appliances such as TV, fan, lighting, doors and change channels, temperature, and volume by just hand gestures.