Distributed Super Resolution Image Processing on Cloud
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional multiple frame super resolution processes on cloud platforms face long processing times and compatibility issues with cloud platforms, especially when handling compressed and encoded images, due to inefficient distribution of tasks across nodes.
Innovation Solution
An information processor divides input image data into multiple frames and allocates them to worker nodes based on determined division points, using a master node to analyze image characteristics and dynamically allocate processing tasks, allowing for efficient parallel processing across multiple nodes while managing node resources and re-division as needed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If multiple frame super resolution is performed as distributed process at plurality of nodes, then processing time is reduced, but compatibility with cloud platform deteriorates and compressed/encoded images cannot be efficiently handled
Solution Approach 1:
The patent segments the image processing task by dividing the input image into multiple blocks and distributing them across different worker nodes. Each worker node processes specific blocks independently, enabling parallel computation that reduces overall processing time while maintaining compatibility with cloud platform architectures through standardized block-based processing interfaces.
Solution Approach 2:
The patent implements a universal processing framework that can handle both compressed and encoded images through the same distributed super resolution pipeline. The system uses a master node that receives various image formats, divides them into blocks, and distributes to worker nodes capable of processing different image types, thereby achieving multi-functionality and improved cloud platform compatibility.
2Adaptability or versatility
If super resolution process is performed at one node, then compatibility with cloud platform is maintained, but processing time becomes too long
Solution Approach 1:
The patent divides the single-node processing task into multiple segments distributed across cloud nodes. The master node coordinates the distribution of image blocks to multiple worker nodes, enabling parallel processing that significantly reduces processing time while maintaining compatibility through standardized cloud computing interfaces and block-based processing.
Solution Approach 2:
The patent transitions from single-node sequential processing to multi-node parallel processing by adding the spatial dimension of distribution across cloud infrastructure. This dimensional change enables simultaneous processing of multiple image blocks across different nodes, reducing processing time while maintaining cloud platform compatibility through standardized networking and data transfer protocols.
3Productivity
If distributed process is implemented without proper division strategy, then processing can be distributed, but node resources cannot be efficiently managed and idle nodes occur
Solution Approach 1:
The patent implements dynamic task distribution where the master node continuously monitors worker node status and adjusts block allocation in real-time. When workers complete their tasks, the system dynamically redistributes available blocks to keep nodes active, preventing idle time and optimizing resource utilization across the distributed cloud infrastructure.
Solution Approach 2:
The patent incorporates feedback mechanisms where worker nodes report completion status to the master node, which then adjusts the distribution strategy accordingly. This feedback loop enables efficient resource management by redirecting blocks to available workers and preventing node idleness, thereby optimizing overall processing productivity.
Data Source
AI summary
According to one embodiment, an information processor divides input image data for a plurality of information processors so as to perform a super resolution process as a distributed processing. The super resolution process generates a high resolution image based on a plurality of low resolution input images. The information processor includes a division point determining unit, a division processor, and a combination processor. The division point determining unit determines a division point for an input image of a sequence of image frames. The division processor divides, for each of the information processors, the input image at the division point, and allocates each divided input image to each corresponding one of the information processors.