Hierarchical Image Processing Latency Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional hierarchical image processing techniques suffer from long processing periods and high latency, especially when applied to live-view displays, due to the need to wait for all necessary pixel lines to be prepared, which hinders real-time image processing.
Innovation Solution
An image processing apparatus and method that generate multiple reduced images with different reduction rates from an input image, using a storage unit to store previous reduced images and a selection unit to output either newly generated or stored reduced images for processing, allowing for composite data creation and efficient hierarchical processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If hierarchical image processing is performed by waiting for all necessary pixel lines to be prepared, then processing accuracy is improved, but processing time increases significantly
Solution Approach 1:
The patent applies preliminary action by generating and storing reduced images in advance before they are actually needed for processing. The storage unit holds previously generated reduced images that can be immediately used when processing is required, eliminating the waiting time for image generation while maintaining processing accuracy. This is evident in the configuration where reduced images are pre-generated and stored, then retrieved and enlarged as needed for composite data creation.
2Manufacturing precision
If multiple reduced images are generated and processed in sequence, then image quality is maintained, but processing latency increases
Solution Approach 1:
The patent transitions from sequential processing to parallel processing by operating across multiple dimensions of image data. Multiple reduced images with different reduction rates are generated and stored simultaneously, then retrieved and processed in parallel to create composite data. This dimensional approach allows image quality to be maintained through multiple processing levels while dramatically reducing processing latency through concurrent operations.
Solution Approach 2:
By pre-generating and storing multiple reduced images at different reduction rates before processing is needed, the system eliminates the sequential generation bottleneck. These pre-prepared images are then retrieved and enlarged simultaneously to create composite data, maintaining image quality across multiple levels while reducing overall processing latency.
3Measurement precision
If all reduced images are generated from the current input image, then processing accuracy is maximized, but real-time processing capability is reduced
Solution Approach 1:
The system performs preliminary action by generating reduced images from previous input images and storing them in advance. When real-time processing is required, these pre-generated reduced images are retrieved and enlarged immediately, eliminating the time-consuming image generation step while maintaining processing accuracy through the use of previously computed data.
4Productivity
If hierarchical processing is performed in units of image blocks with pipeline processing, then processing throughput is improved, but processing period for each image increases
Solution Approach 1:
By pre-generating and storing reduced images before processing is needed, the system eliminates the sequential waiting period that occurs in pipeline processing. Multiple reduced images are prepared in advance and can be retrieved simultaneously for composite data creation, maintaining high processing throughput while significantly reducing the processing period for each individual image.
Data Source
AI summary
Disclosed is an image processing apparatus that generate a plurality of reduced images with different reduction rates from an input image, applies predetermined processing to the reduced images, enlarges the reduced images into their respective original resolutions, and then composes the enlarged images into a composite image. The apparatus applies processing to the input image based on the composite image. The apparatus, depending on a setting, applies the processing to an input image based on the composite image generated from a previous input image.


