Real-time Double hand Interaction for 8k 3D Light Field Display Based on implicit neural network

• 1、School of Electronic Engineering,Beijing University of Posts and Telecommunications

Abstract：In this paper, we present a real-time 3D display interaction system. The system enables users to accurately interact with 3D objects in real-time display of 3D light fields. Our double-handed gesture recognition module obtains skeleton information of both hands in real-time and processes key point coordinates by normalization and positional encoding, so gesture information is recognized using the MLP network. The 3D rendering module processes hand gesture and displacement information in real-time, and jointly analyzes 3D rendering scene information to enable real-time interaction of the light field. The results of our directional quantitative experiments show that our proposed method achieves real-time double-handed interaction for different displays and various 3D rendering engines. Our proposed method offers better interaction speed and accuracy than previous methods. It effectively overcomes the limitations of occlusion and slow speed that are often associated with two-handed interaction and also improves the relatively single interaction mode of the 3D light field display. This technology\'s potential for real-world applications is significant, making it a valuable contribution to the development of 3D light field display interaction.

Keywords： Artificial intelligence, real-time interaction with two hands, three-dimensional light field display.