finger type classification from depth image

This traditional approach has high recognition accuracy, but is costly and poorly scalable and easy to use. Traditional gesture recognition requires the use of data gloves 5 or other relevant external devices to collect the spatial position changes of the hand and arm joints to determine the real intention of the wearer. In the field of computer vision, gesture recognition is also one of the most important and topical problems and has been used in many fields, such as human–computer interaction 1, virtual reality systems 2, 3, and sign language recognition 4. In addition to the use of language, human beings use gestures as an indispensable communication tool when communicating and conveying information. The experimental results show that our algorithm has a high accuracy while ensuring low network complexity. Finally, we use Cambridge Hand Gesture dataset and Northwestern University Hand Gesture dataset to verify the effectiveness of our algorithm. To ensure the quality of the extracted optical flow graph without increasing the complexity of the network, we use the fractional-order method to extract the optical flow graph, creatively combine fractional calculus and deep learning. The original keyframes and optical flow keyframes are used to represent spatial and temporal features respectively, which are then sent to the 2D convolutional neural network for feature fusion and final recognition. To improve this problem, we propose a recognition method based on a strategy combining 2D convolutional neural networks with feature fusion.

Although 3D convolutional neural networks consider both spatial and temporal features, the networks are too complex, which is the main reason for the low efficiency of the algorithms.

Currently, most of the commonly used dynamic gesture recognition algorithms are based on 3D convolutional neural networks. Therefore, proposing a dynamic gesture recognition algorithm that balances efficiency and accuracy is still a meaningful work.

In recent years, many algorithms for gesture recognition have been proposed, but most of them do not have a good balance between recognition efficiency and accuracy. Gesture recognition is one of the most popular techniques in the field of computer vision today.

Finger type classification from depth image