Vector Processor and Memory Fabric for Low Power Computational Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computational imaging technologies face challenges in adoption due to high computational resource requirements and power consumption, particularly in portable devices, where they need to process high-resolution images and videos quickly while maintaining low power dissipation and latency.
Innovation Solution
A computing device with vector processors, hardware accelerators, and a memory fabric that enables efficient processing of images and videos by providing a flexible infrastructure with power management modules to optimize power usage and performance, allowing for customizable processing pipelines and efficient communication between components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If computational imaging processes high-resolution images and videos quickly, then processing speed is improved, but power consumption increases
Solution Approach 1:
The system segments the computational imaging pipeline into multiple processing stages (image acquisition, preprocessing, computational processing, output generation) and assigns different processing modes to different stages. Vector processors handle computationally intensive tasks while hardware accelerators handle specific operations, allowing the system to optimize power consumption at each segment while maintaining overall processing speed.
Solution Approach 2:
The system dynamically adjusts processing resources based on input characteristics and required output quality. The flexible hardware architecture allows real-time reconfiguration of processing pipelines, enabling the system to use high-performance modes only when necessary and low-power modes during routine operations, thus resolving the contradiction between speed and power consumption.
2Productivity
If computational imaging uses high computational resources, then processing capability is improved, but device complexity increases
Solution Approach 1:
The patent implements universal processing units that can perform multiple functions. Vector processors can handle various computational imaging algorithms (depth mapping, panoramic stitching, object recognition) while hardware accelerators provide specialized support for common operations. This multi-functionality reduces the need for dedicated hardware for each task, thereby reducing overall device complexity while maintaining high processing capability.
Solution Approach 2:
The system introduces a flexible hardware infrastructure layer that acts as an intermediary between the image sensor and the application layer. This infrastructure provides standardized interfaces and processing pipelines that simplify the integration of different computational imaging algorithms, reducing the complexity burden on both the hardware design and software implementation.
3Loss of time
If computational imaging reduces latency, then user experience is improved, but power consumption increases
Solution Approach 1:
The system performs preliminary processing operations in advance to reduce latency during critical operations. Hardware accelerators pre-process images and videos using dedicated circuits that operate at lower power than general-purpose processors. By preparing processed data in advance and maintaining ready-state processing pipelines, the system achieves low latency without continuously consuming high power.
Solution Approach 2:
The patent applies different processing qualities to different regions or frames based on their importance. Critical regions requiring low latency (such as focus areas or motion-detecting zones) receive accelerated processing through hardware accelerators, while less critical regions use lower-power processing modes. This localized quality approach reduces overall power consumption while maintaining low latency where it matters most.
Data Source
AI summary
The present application discloses a computing device that can provide a low-power, highly capable computing platform for computational imaging. The computing device can include one or more processing units, for example one or more vector processors and one or more hardware accelerators, an intelligent memory fabric, a peripheral device, and a power management module. The computing device can communicate with external devices, such as one or more image sensors, an accelerometer, a gyroscope, or any other suitable sensor devices.