Odd-Group Pixel Down-Sampling for Saw-Tooth Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image down-sampling methods in surveillance cameras often result in saw-tooth images due to non-uniform sampling or aliasing, leading to incorrect event detection.
Innovation Solution
An image down-sampling method that groups pixels in odd numbers to generate output pixel values, addressing the issue by performing down-sampling using groups of odd pixels to produce output images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If conventional down-sampling methods are used to reduce power consumption, then power consumption is reduced, but saw-tooth images are generated due to non-uniform sampling or aliasing
Solution Approach 1:
The patent divides the pixel array into multiple down-sampling groups, where each group contains a specific number of pixels (including odd numbers) that are processed together to generate output pixel values. This segmentation approach allows controlled down-sampling that prevents saw-tooth artifacts while maintaining power savings.
Solution Approach 2:
The patent changes the down-sampling parameters by using groups with odd numbers of pixels instead of conventional even-numbered groups or simple uniform sampling. This parameter change in the grouping strategy eliminates non-uniform sampling artifacts and aliasing while still achieving reduced power consumption.
2Use of energy by moving object
If down-sampling is performed to generate images for event detection, then power consumption is decreased, but saw-tooth images cause wrong event decisions
Solution Approach 1:
The patent segments the pixel array into down-sampling groups with odd numbers of pixels, ensuring that the down-sampled images maintain sufficient detail for accurate event detection. This segmentation prevents the saw-tooth artifacts that would otherwise lead to wrong event decisions.
Solution Approach 2:
By changing the down-sampling parameter to use odd-numbered pixel groups, the patent maintains image quality necessary for reliable event detection while achieving power savings. This parameter change ensures that down-sampled images do not contain artifacts that would cause detection errors.
3Productivity
If uniform down-sampling is applied to reduce computational complexity, then processing speed is improved, but aliasing and non-uniform sampling artifacts are introduced
Solution Approach 1:
The patent segments pixels into groups with odd numbers of elements, creating a non-uniform down-sampling pattern that prevents aliasing. This segmented approach maintains processing efficiency while improving image quality by eliminating artifacts.
Solution Approach 2:
The patent introduces asymmetry in the down-sampling groups by using odd numbers of pixels instead of symmetric even numbers. This asymmetric grouping strategy breaks the patterns that cause aliasing and saw-tooth artifacts while maintaining computational efficiency.
Data Source
AI summary
An image down-sampling method, for down-sampling a sensing image comprising a plurality of first pixels and a plurality of second pixels which are alternatively arranged, comprising: generating a first output pixel value according to at least one pixel value of at least one of the first pixel in a first down-sampling group; wherein the first down-sampling group comprises an odd number of pixels. By this way, the saw-tooth image issue in prior art may be improved. If the image down-sampling method is applied to a surveillance camera, the accuracy of event detection of the surveillance camera may be increased.