Heterogeneous Hand Gesture Recognition using 3D Dynamic Skeletal Data
C. Anitha · International Journal for Research in Applied Science and Engineering Technology · 2021
Hand gestures area unite the foremost natural and intuitive non-verbal communication medium whereas interacting with a pc, and connected analysis have recently boosted interest. To boot, the distinctive options of the hand provided by current business cheap depth -camera are often exploited in numerous gesture recognition based systems, for human -computer interaction.This paper builds a sturdy hand form options from two modalities of depth and skeleton form approach, we have a tendency to use the movements, the rotations of the hand joints with relevance to their neighbors, and also the skeleton point cloud to find out the 3D geometric transformation.For the hand depth form approach, we have tendency to use the feature illustration from the hand element segmentation model. Finally, we propose a multi-level feature LSTM with CONV1D and CONV2D algorithm where LSTM is used to manage the range of hand options. Therefore, we tend to propose a completely unique technique by exploiting skeletal point clouds from skeletal form as well depth features from real hand depth form in order for the LSTM model to benefit from both. Our projected technique achieves the best result with skeletal and depth data.