<![CDATA[In this paper, we propose and implement a novel and real-time method for recognizing hand gestures using depth map. The depth map contains information relating to the distance of objects from a viewpoint. Microsoft’s Kinect sensor is used as input device to capture both the color image and its corresponding depth map. We first detect bare hand in cluttered background using distinct gray-level of the hand which is located near to the sensor. Then, scale invariance feature transform (SIFT) algorithm is used to extract feature vectors. Lastly, vocabulary tree along with K-means clustering method are used to partition the hand postures to ten simple sets as: "one", "two", "three", "four", "five", "six", "seven", "eight", "nine" and "ten" numbers based on the number of extended fingers. The vocabulary tree allows a larger and more discriminatory vocabulary to be used efficiently. Consequently, it leads to an improvement in accuracy of the clustering. The experimental results show superiority of the proposed method over other available approaches. With this approach, we are able to recognize 'numbers' gestures with over 90% accuracy.DOI:http://dx.doi.org/10.11591/ijece.v3i6.3954]]>
Copyrights © 2013
