Image Processing Apparatus Memory Reduction via Multi-Level Smoothing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional low-pass filtering for image smoothing requires large memory capacity and fails to accurately reproduce edges due to loss of fine image information during the reduction process.
Innovation Solution
An image processing apparatus that generates smoothed images by reducing the memory capacity needed for LPF processing by holding intermediate smoothed images at multiple preset level values, allowing for accurate edge reproduction by synthesizing these images with subsequent frames without performing additional reduction on them.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If image reduction process is performed for smoothing, then memory capacity required is reduced, but edge accuracy deteriorates
Solution Approach 1:
The patent segments the smoothing process into multiple stages by creating intermediate smoothed images at different reduction rates (e.g., 1/2, 1/4, 1/8). Each intermediate image preserves different levels of detail, allowing the system to select appropriate segments for synthesis. This segmentation enables memory reduction while maintaining edge accuracy by combining results from multiple segmentation levels.
Solution Approach 2:
The patent introduces a new dimension of processing by creating multiple intermediate smoothed images at different reduction rates rather than performing a single reduction. This multi-dimensional approach allows synthesis of results from different reduction levels, effectively adding a dimensional layer to the smoothing process that preserves edge information while reducing memory requirements.
2Reliability
If conventional LPF process is performed, then smoothing effect is achieved, but memory capacity requirement increases
Solution Approach 1:
The patent applies partial reduction actions at multiple stages rather than a single full reduction. By performing smoothing operations at intermediate reduction rates (partial actions) and synthesizing the results, the system achieves the desired smoothing quality without requiring memory capacity for complete full-resolution processing.
3Productivity
If image reduction is performed, then processing speed may improve, but loss of fine information increases
Solution Approach 1:
The patent performs preliminary smoothing actions at multiple reduction rates before final synthesis. By pre-computing intermediate smoothed images at different reduction levels, the system prepares processed data in advance that retains essential fine information, enabling faster final processing without complete information loss.
Solution Approach 2:
The patent changes the reduction rate parameter across multiple processing stages, creating intermediate images at different reduction levels (e.g., 1/2, 1/4, 1/8). This parameter variation allows the system to process images at different resolutions and synthesize results that preserve fine information while benefiting from the speed advantages of reduced-resolution processing.
Data Source
AI summary
An image processing apparatus holds a plurality of intermediate smoothed images smoothed at a plurality of preset level values, reduces the N-th frame image of a moving image received from the outside to generate a reduced image, performs a smoothing process on the generated reduced image at the plurality of preset level values to generate a plurality of intermediate smoothed images, stores the generated intermediate smoothed images in an intermediate smoothed image storing unit, acquires, when generating the smoothed images of the frames after the (N+1)-th frame, one or a plurality of intermediate smoothed images from the plurality of intermediate smoothed images of the N-th frame stored in the intermediate smoothed image storing unit, synthesizes the acquired intermediate smoothed images of the N-th frame and the frames after the (N+1)-th frame of the moving image received from the outside, and generates smoothed images.