Fuzzy Control
User adaptive handwriting recognition by self-growing probabilistic decision-based neural networks
IEEE Transactions on Neural Networks
A flexible way of modeling the long-term cost competitiveness of a semiconductor product
Robotics and Computer-Integrated Manufacturing
| Hi-index | 0.00 |
Thanks to the rapid advancement of human-computer interaction technologies it is becoming easier for the elderly and/or people with disabilities to operate various electrical systems. Operation of home appliances by using a set of predefined hand gestures is an example. However, hand gesture recognition may fail when the predefined command gestures are similar to some ordinary but meaningless behaviors of the user. This paper uses a gesture spotting method to recognize a designated gesture from other similar gestures. A fuzzy garbage model is proposed to provide a variable reference value to determine whether the user's gesture is the command gesture or not. Further, the authors propose two-stage user adaptation to enhance recognition performance: that is, off-line batch adaptation for inter-person variation and on-line incremental adaptation for intra-person variation. For implementation of the two-stage adaptation method, a genetic algorithm GA and the steepest descent method are adopted for each stage. Experimental results were obtained for 5 different users with left and up command gestures.