Reconfigurable Depth Estimation Circuit for Video Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems lack an efficient and resource-effective solution for real-time depth estimation of both stereoscopic and monoscopic video data, as they require dedicated circuits for each type, leading to hardware redundancy and inefficiency.
Innovation Solution
A system-on-chip with reconfigurable shared circuits that can switch between monoscopic and stereoscopic depth estimation modes, using an input processing circuit for downsampling and a depth upscaling circuit for upsampling, allowing the same hardware to handle both types of data with different resolution requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If dedicated circuits are implemented for both monoscopic and stereoscopic depth estimation, then both depth estimation functions can be performed, but hardware redundancy increases and resource efficiency decreases
Solution Approach 1:
The patent implements a single reconfigurable depth estimation circuit that can function as either a monoscopic depth estimator or a stereoscopic depth estimator based on configuration data. The circuit includes reconfigurable shared components such as an input processing circuit that can spatially downscale video data in different modes, and a depth upscaling circuit that can upscale depth data according to different requirements. This universal circuit design eliminates the need for separate dedicated circuits for each depth estimation type, reducing hardware redundancy while maintaining full functionality for both monoscopic and stereoscopic processing
Solution Approach 2:
The depth estimation circuit is designed to be dynamically reconfigurable between different operational modes. Configuration data determines whether the circuit operates in monoscopic mode or stereoscopic mode, allowing the same hardware to adapt its behavior based on the input video type. The input processing circuit can dynamically adjust downscaling parameters, and the depth upscaling circuit can dynamically modify upscaling operations based on the active mode, enabling flexible resource utilization without hardware changes
2Reliability
If separate depth estimators are used for monoscopic and stereoscopic video data, then each estimator can be optimized for its specific use case, but device complexity increases
Solution Approach 1:
The patent employs a universal reconfigurable depth estimation circuit that maintains high reliability for both monoscopic and stereoscopic processing through mode-specific optimization. The circuit receives configuration data indicating the active mode and adjusts its operation accordingly - using monoscopic processing paths when handling 2D video data and stereoscopic processing paths when handling 3D video data. This approach achieves use-case-specific optimization without requiring physically separate estimators, thereby maintaining depth estimation accuracy while reducing overall device complexity
Solution Approach 2:
The reconfigurable circuit utilizes parameter changes to adapt between different depth estimation modes. Configuration data modifies operational parameters such as downscaling factors, processing algorithms, and upscaling characteristics based on whether monoscopic or stereoscopic mode is active. This parameter-based reconfiguration allows the same hardware circuit to achieve optimized performance for each specific use case without structural complexity of separate dedicated circuits
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system-on-chip is provided which is configured for real-time depth estimation of video data. The system-on-chip comprises a monoscopic depth estimator (200) configured to perform monoscopic depth estimation from monoscopic-type video data, and a stereoscopic depth estimator (210) configured to perform stereoscopic depth estimation from stereoscopic-type video data. The system-on-chip is reconfigurable to perform either the monoscopic depth estimation or the stereoscopic depth estimation on the basis of configuration data defining a selected depth estimation mode. Both depth estimators comprise shared circuits (120, 150) which are instantiated in hardware and reconfigurable to account for differences in the functionality of the circuit in each depth estimator.