Eye-Coordinate Viewpoint Maps for Lower-Volume 3D Video Communication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional 3D video communication methods require multiple cameras, leading to high hardware costs, data transmission costs, and high power consumption due to the need for multiple GPU hardware encoding/decoding chips and large data transmission requirements.
Innovation Solution
Obtain human eye positioning coordinates at the display terminal, render left-eye and right-eye viewpoint maps based on these coordinates, and transmit these maps to the display terminal, reducing the need for encoding and decoding of full scene data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional multi-camera system with GPU hardware encoding/decoding is used, then 3D video communication quality is maintained, but hardware cost, data transmission cost, and power consumption increase significantly
Solution Approach 1:
The patent extracts and transmits only the essential viewpoint map data required for 3D reconstruction rather than transmitting complete multi-camera scene data. By separating the viewpoint synthesis function from the data transmission, the system reduces hardware requirements while maintaining 3D video quality.
Solution Approach 2:
The patent segments the video data processing into two parts: (1) rendering viewpoint maps at the acquisition terminal using minimal computational resources, and (2) transmitting only these compressed viewpoint maps rather than full scene data. This segmentation reduces the burden on transmission hardware and decoding devices.
2Reliability
If traditional multi-camera system with GPU hardware encoding is used, then complete scene data is captured, but data transmission requirements and network bandwidth increase significantly
Solution Approach 1:
The patent creates a simplified copy of the scene data in the form of viewpoint maps that contain only the essential visual information needed for reconstruction. These viewpoint map copies are then transmitted instead of the original complete scene data, dramatically reducing transmission volume while preserving the ability to reconstruct 3D video.
3Productivity
If multiple GPU hardware encoding/decoding chips are used, then encoding/decoding performance is sufficient, but power consumption increases significantly
Solution Approach 1:
The patent replaces the mechanical GPU hardware encoding/decoding system with a software-based viewpoint map rendering and synthesis approach. By substituting hardware-intensive operations with algorithmic processing, the system achieves comparable performance with significantly reduced power consumption.
Data Source
AI summary
The present application proposes a video communication method, including obtaining human eye positioning coordinate data of a viewer acquired at a display terminal, wherein the human eye positioning coordinate data includes a horizontal coordinate of the left eye and a horizontal coordinate of the right eye of the viewer in the display space of the display terminal, acquiring a current frame scene image of a scene located at the acquisition terminal, rendering a left-eye viewpoint map at a viewpoint corresponding to the horizontal coordinate of the left eye and a right-eye viewpoint map at a viewpoint corresponding to the horizontal coordinate of the right eye in the display space according to the current frame scene image and the human eye positioning coordinate data, transmitting rendered left-eye viewpoint map and right-eye viewpoint map to the display terminal so as to perform display at the display terminal.