Image Processing Apparatus Region-Based Blur Kernel Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for adding blur to images require significant calculations, especially when dealing with continuous distance changes, leading to increased calculation time and unnatural blur effects when attempting to reduce calculation amounts by limiting object sides or treating continuous changes as discontinuous data.
Innovation Solution
An image processing apparatus that divides a shot image into regions with controlled image quality changes, calculates a blur kernel for each region, and generates a blur-added image through convolution calculations, reducing overall calculation requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If blur is added to image data containing continuous distance changes using a blur kernel depending on object distance, then natural blur effect is achieved, but calculation amount increases significantly
Solution Approach 1:
The image is divided into multiple regions based on depth of field information, where each region corresponds to a specific depth range. A single blur kernel is applied to each region instead of calculating different kernels for every pixel, significantly reducing calculation amount while maintaining natural blur effects.
Solution Approach 2:
Different blur kernels are applied to different regions of the image based on their depth characteristics. Each region receives a blur kernel appropriate to its depth range, ensuring local blur quality matches the expected depth of field effect without requiring complex per-pixel calculations.
2Loss of time
If GPU hardware acceleration is used for blur calculation, then calculation time is decreased, but calculation amount remains large
Solution Approach 1:
By segmenting the image into depth-based regions, the total calculation amount is reduced from per-pixel kernel calculations to regional kernel applications. This reduces the computational workload that would otherwise require GPU acceleration, making the process more efficient.
3Quantity of substance
If object side limits are applied or continuous distance changes are treated as discontinuous data to reduce calculation, then calculation amount is reduced, but image loses sense of distance or blur becomes unnatural
Solution Approach 1:
The image is segmented into depth regions that preserve continuous distance information through proper region boundaries. This segmentation maintains the sense of distance by ensuring regions correspond to meaningful depth transitions, avoiding the unnatural effects of arbitrary discontinuous treatment.
Solution Approach 2:
Each depth region is assigned a blur kernel that locally matches its depth characteristics, preserving the natural appearance of blur at different distances. This local quality approach maintains image realism without requiring exhaustive per-pixel calculations.
Data Source
AI summary
An image processing apparatus 1 includes an image region dividing portion 105 configured to divide a shot image into a plurality of image regions so that an amount of change of an image quality is within a predetermined range when the shot image is blurred using a blur kernel depending on an object distance and an angle of view of the shot image, a blur kernel generating portion 106 configured to calculate the blur kernel for each image region divided by the image region dividing portion 105, and an image processing calculating portion 107 configured to generate a blur-added image by performing a convolution calculation for the shot image using the blur kernel calculated by the blur kernel generating portion 106.