Edge Extraction for Shake Blur Correction in Digital Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for correcting blurred images, particularly those caused by shake blur in cell phone cameras, are inefficient due to the need for specialized devices to measure blur direction and width, and repetitive processes that consume processing time, making it difficult to apply these methods to cell phone cameras and resulting in inaccurate or delayed detection of blur.
Innovation Solution
An image processing method that determines optimal reduction intensities and scanning intervals based on image size to efficiently extract edge data from digital photographic images, allowing for blur correction without the need for specialized devices and reducing processing time by analyzing edge states to determine blur direction and width.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If edges are extracted from digital photographic images at original size, then edge extraction accuracy is improved, but processing time increases significantly
Solution Approach 1:
The patent segments the image processing task by extracting edges from reduced-resolution images rather than full-resolution images. This segmentation of the processing workload allows for faster computation while maintaining sufficient edge detection accuracy for blur correction purposes.
Solution Approach 2:
The patent changes the resolution parameter of the input image before edge extraction. By reducing the image resolution to a lower resolution image, the system processes fewer pixels while still capturing the essential edge information needed for determining blur characteristics.
2Loss of time
If reduction intensity is increased for large images, then processing time is reduced, but edge extraction accuracy may deteriorate
Solution Approach 1:
The patent dynamically adjusts the reduction intensity based on the original image size. Different reduction ratios are applied to different image categories (e.g., 1/2 for small images, 1/4 for medium images, 1/8 for large images), optimizing the balance between processing speed and edge detection accuracy for each image size.
3Manufacturing precision
If repetitive correction and evaluation processes are used, then image quality is improved, but processing time increases significantly
Solution Approach 1:
The patent performs preliminary edge extraction from reduced-resolution images to obtain edge data that represents the states of edges. This preliminary action provides sufficient blur information without requiring repetitive correction and evaluation cycles, significantly reducing processing time while maintaining adequate image quality.
Solution Approach 2:
The patent extracts only the essential edge information from the image using reduced-resolution processing. By taking out only the necessary edge data rather than processing the entire image through multiple iterative corrections, the system achieves sufficient blur correction without the time cost of repetitive processes.
4Measurement precision
If specialized devices are provided to obtain blur data, then blur detection accuracy is improved, but device complexity increases
Solution Approach 1:
The patent enables the digital camera system to self-determine blur characteristics by analyzing edge states from the captured image itself. The camera uses its existing image processing capabilities to extract edge data and calculate blur direction and width, eliminating the need for external specialized devices like accelerometers while maintaining accurate blur detection.
Solution Approach 2:
The patent uses edge data as an intermediary to infer blur characteristics. Instead of directly measuring blur with specialized hardware, the system analyzes the states of edges extracted from the image, which serve as an intermediary representation that contains information about the blur conditions without requiring additional sensing devices.
Data Source
AI summary
Edges are efficiently extracted from digital photographic images. A reduction rate determining means determines reduction rates for images, such that larger images are assigned higher reduction intensities. A reduction executing means reduces an image, employing the reduction rate determined by the reduction rate determining means, to obtain a reduced image. An edge detecting means extracts edges in eight directions from the reduced image. Coordinate positions of the edges in each directions are obtained, and output to an edge profile generating means.