Visual tracking of high DOF articulated structures: an application to human hand tracking
ECCV '94 Proceedings of the third European conference on Computer Vision (Vol. II)
Real-Time American Sign Language Recognition Using Desk and Wearable Computer Based Video
IEEE Transactions on Pattern Analysis and Machine Intelligence
Extraction of 2D Motion Trajectories and Its Application to Hand Gesture Recognition
IEEE Transactions on Pattern Analysis and Machine Intelligence
ECCV '98 Proceedings of the 5th European Conference on Computer Vision-Volume II - Volume II
An Appearance-Based Approach to Gesture-Recognition
ICIAP '97 Proceedings of the 9th International Conference on Image Analysis and Processing-Volume II
GREFIT: Visual Recognition of Hand Postures
GW '99 Proceedings of the International Gesture Workshop on Gesture-Based Communication in Human-Computer Interaction
Robust classification of hand postures against complex backgrounds
FG '96 Proceedings of the 2nd International Conference on Automatic Face and Gesture Recognition (FG '96)
Towards 3D hand tracking using a deformable model
FG '96 Proceedings of the 2nd International Conference on Automatic Face and Gesture Recognition (FG '96)
View-Based Active Appearance Models
FG '00 Proceedings of the Fourth IEEE International Conference on Automatic Face and Gesture Recognition 2000
Bare-hand human-computer interaction
Proceedings of the 2001 workshop on Perceptive user interfaces
Automatic Sign Language Analysis: A Survey and the Future beyond Lexical Meaning
IEEE Transactions on Pattern Analysis and Machine Intelligence
Visual hand posture recognition in monocular image sequences
DAGM'06 Proceedings of the 28th conference on Pattern Recognition
The Journal of Machine Learning Research
| Hi-index | 0.00 |
We propose a novel method for extracting natural hand parametersfrom monocular image sequences. The purpose is to improve avision-based sign language recognition system by providing detailinformation about the finger constellation and the 3D hand posture.There for the hand is modelled by a set of 2D appearance models,each representing a limited variation range of 3D hand shape andposture. The single models are linked to each other according tothe natural neighbourhood of the corresponding hand status. Duringan image sequence, necessary model transitions are executed towardsone of the current neighbour models. The natural hand parametersare calculated from the shape and texture parameters of the currentmodel, using a relation estimated by linear regression. The methodis robust against large differences between subsequent frames andalso against poor image quality. It can be implemented in real-timeand offers good properties to handle occlusion and partly missingimage information.