Democritus University of Thrace (DUTH)
Ανάπτυξη τεχνικών οπτικής αναγνώρισης χειρονομιών
Abstract
dc:descriptionA series of vision based hand gesture recognition techniques have been developed in the context of this thesis concerning: (i) hand detection in color images and videos, (ii)estimation of hand posture formed by raised fingers, (iii) dynamic gesture recognition, and (iv) body parts localization using depth imaging. There were developed two techniques for hand detection in color images and videos which address the difficulties associated with the variety of cameras and lighting conditions and the complexity of the background. The first hand detection technique uses a skin color filter which combines a dynamic skin color model and an explicitly defined skin color region. The second technique employs combination of existing algorithms for motion detection and introduces a novel skin color classifier to improve segmentation accuracy. Motion detection is based on image differencing and background subtraction. Skin color detection is accomplished via a color classification technique that employs online color training, so that the system can dynamically adapt to the variety of lighting conditions and the user's skin color in the most effective way. Morphological features of the detected hand in previous frames are employed to estimate the probability of a pixel belonging to the hand section in the current frame. Finally, the derived motion, color and morphological information are combined to detect the hand region. Hand pose estimation is accomplished via hand skeleton extraction which is based on horizontal distance transform of the hand’s image. Dynamic gesture recognition is achieved by the classification of hand’s trajectory features. The classifier is based on Hidden Markov Models. The difficulties faced in hand detection in intensity images and videos along with the recent advance in depth sensing technology led to the development of an upper body parts localization technique using depth imaging. Based on the depth map provided by the Microsoft Kinect device we are able to detect six parts of the body using the self-organized and self-growing neural network SGONG. Then, hand region extraction is used for static gesture recognition while the hand’s position is tracked by using Kalman filtering for dynamic gesture recognition. Lastly, a digital map display software has been developed, which is controlled by static and dynamic gestures and exploits the three-dimensional localization of the hand using depth imaging.
Degree
thesis:*- Grantor dc:publisher
- Democritus University of Thrace (DUTH)
- Year dc:date
- 2014
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Sgouropoulos, Kyriakos
- Σγουρόπουλος, Κυριάκος
Subjects
dc:subject × 14- Αναγνώριση χειρονομιών
- Αναγνώριση χειρομορφών
- Όραση βάθους
- Ανίχνευση χεριού
- Εκτίμηση χειρομορφής
- Gesture recognition
- Hand detection
- Depth imaging
- Hand posture
- Posture estimation
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/35176