An implementation of sign language alphabet hand posture recognition using geometrical features through artificial neural network (part 2)

摘要: In the sign language alphabet, several hand signs are in use. Automatic recognition of performed can facilitate communication between hearing and none people. This framework proposes posture American Sign Language alphabet based on a neural network (NN) which works geometrical feature extraction hand. The user’s is captured by 3D depth-based sensor camera. Consequently, segmented according to depth features. proposed system called ‘Depth-based Geometrical Recognition’ (DGSLR). DGSLR adopted an easier segmentation approach, further used other applications. improves accuracy due unchangeable features against orientation or rotation compared Discrete Cosine Transform (DCT) Moment Invariant. As support vector machine (SVM) type artificial (ANN), it drive desired outcomes. Since there 26 different multi-class SVM versus single classifier with classes RBF kernel was validate each class. proficient provides up 96.78%. findings iterations demonstrated that combination extracted resulted better rate process classification step.