Image Processing Apparatus Dynamic Color Component Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing methods face challenges with increased circuit scale and memory usage as the size of input images grows, and require proportional increases in the number of ports and terminals on PCIe switches when adding processing modules.
Innovation Solution
An image processing apparatus and method that dynamically allocates color components for processing between multiple modules based on communication amounts, optimizing the number of color components processed by each module to reduce the overall circuit scale and data transfer load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple image processing modules are connected in parallel via PCIe switch to improve processing speed, then productivity increases, but device complexity increases due to proportional increase in number of ports and terminals
Solution Approach 1:
The image data processing is segmented by color components (e.g., Y, Cb, Cr channels separated). Each processing module handles specific color components rather than all components, reducing the data volume each module must process and transmit. This segmentation allows parallel processing without proportionally increasing circuit scale, as each module has a dedicated, reduced data path.
Solution Approach 2:
Different processing modules are assigned different color component processing tasks based on their capabilities and the specific processing requirements. The controller dynamically determines which modules process which color components, optimizing the distribution of processing loads and minimizing the circuit resources required for data transfer between modules and controller.
2Device complexity
If multiple image processing modules are connected in series to process image data, then device complexity is reduced, but productivity decreases due to sequential processing
Solution Approach 1:
The processing pipeline is segmented such that different color components are processed in parallel by different modules rather than sequentially. Multiple modules can simultaneously process different color channels of the same image data, achieving parallel processing without requiring a complex mesh network of connections between all modules.
Solution Approach 2:
The patent introduces a dimensional organization where processing modules are arranged in a hierarchical structure with the controller at the top level and modules at the processing level. Data flows in a structured pattern (controller → module → controller → next module) rather than requiring direct peer-to-peer connections, reducing the overall circuit complexity while enabling parallel processing through the controller's coordination.
3Measurement precision
If image data size increases to improve image quality, then measurement precision improves, but loss of time increases due to longer memory usage and transfer time
Solution Approach 1:
Large image data is segmented by color components, with each segment processed separately by different modules. This segmentation reduces the amount of data that needs to be transferred through any single data path, decreasing transfer time while maintaining full image quality through comprehensive processing of all color components across multiple modules.
Solution Approach 2:
While one processing module is processing image data, the controller simultaneously prepares and transfers data for the next module. The pipeline maintains continuous operation with overlapping processing and transfer operations, reducing the total time required to process large image files while maintaining high image quality through complete processing of all color components.
Data Source
AI summary
A first processing module receives image data transferred from a controller, and executes image processing and a second processing module receives the image data transferred from the controller via the first processing module and executes image processing on the image data. The controller determines color components, of the plurality of color components of the image data, to be processed by the first processing module and the second processing module, based on at least communication amounts produced between the first processing module and the second processing module. Processed data having undergone the image processing performed by the first processing module based on the determined color components is transferred to the controller, and processed data having undergone the image processing performed by the second processing module based on the determined color components is transferred to the controller via the first processing module.