Image Processing Filter Acquisition Using Gain Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing image filtering processes face challenges with increasing storage requirements due to the exponential growth of filter data with varying imaging conditions and limited hardware resources, leading to high calculation costs for determining filter coefficients, especially when performing image filtering with a limited number of taps.
Innovation Solution
An image processing apparatus that performs multiple filtering processes with gain adjustment based on target frequency characteristics, using a filtering process unit that applies filters and gains to original image data, and a gain specifying unit that selects gains from a gain table based on optical characteristics, reducing the data amount and calculation resources needed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If filters are prepared for all combinations of imaging conditions, then filtering accuracy is improved, but storage capacity requirements increase exponentially
Solution Approach 1:
The patent segments the filtering process into two distinct stages: a first filtering process that applies a filter to the input image, and a second filtering process that applies another filter to the result of the first filtering. This segmentation allows the system to achieve high-accuracy filtering equivalent to using a single complex filter while using multiple simpler filters with fewer taps, thereby reducing storage capacity requirements.
Solution Approach 2:
The patent introduces dynamic gain adjustment where gain values are applied to the filtering processes based on frequency characteristics. The gain specifying unit dynamically selects gain values from a gain table based on the imaging conditions and frequency characteristics of the input image, allowing the filtering system to adapt to different conditions without requiring pre-computed filters for all possible conditions.
2Measurement precision
If the number of taps in filters is increased, then filtering accuracy is improved, but calculation amount increases
Solution Approach 1:
The patent divides a single high-order filtering operation into multiple lower-order filtering operations. Instead of using one filter with many taps, the system uses a sequence of filters with fewer taps each, performing multiple filtering passes. This reduces the computational complexity of each individual filtering operation while achieving the same overall filtering accuracy.
Solution Approach 2:
The patent employs periodic action by performing multiple filtering processes in sequence, with each process contributing to the final filtering result. The gain adjustment is also applied periodically based on frequency characteristics, allowing the system to achieve accurate filtering through repeated application of simpler operations rather than a single complex operation.
3Measurement precision
If multiple filtering processes are performed, then filtering accuracy is improved, but processing time increases
Solution Approach 1:
The patent segments the filtering task into multiple specialized filtering processes, each optimized for specific frequency characteristics. The first filtering process handles certain frequency components while the second filtering process handles others, with gain adjustment optimizing the contribution of each process. This segmentation allows parallel processing potential and reduces the computational burden of each individual process.
Solution Approach 2:
The patent changes parameters dynamically by adjusting gain values based on frequency characteristics and imaging conditions. The gain specifying unit selects appropriate gain values from a pre-computed gain table, allowing the filtering system to adapt to different input images and conditions without requiring re-computation of filter coefficients, thereby reducing processing time.
4Measurement precision
If filters are customized for specific frequency characteristics, then filtering accuracy is improved, but device complexity increases
Solution Approach 1:
The patent creates a universal filtering system that can handle multiple frequency characteristics through a standardized two-stage filtering process with gain adjustment. Instead of requiring separate specialized filters for each frequency characteristic, the system uses a single flexible framework that adapts to different frequency requirements through gain specification, reducing device complexity.
Solution Approach 2:
The patent achieves frequency-specific filtering by changing gain parameters rather than changing the filter structures themselves. The gain specifying unit adjusts the gain values applied to the filtering processes based on the frequency characteristics of the input image and imaging conditions, allowing the same filtering hardware to efficiently handle different frequency requirements without structural modifications.
Data Source
AI summary
Further, there are provided a filter acquisition apparatus, a filter acquisition method, a program, and a recording medium, capable of acquiring a filter which is suitably usable in such a filtering process. An image processing apparatus 35 includes a filtering process unit 41 that performs an image filtering process that has a plurality of times of filtering processes. The filtering process unit 41 applies a filter to processing target data to acquire filter application process data, applies a gain to the filter application process data to acquire gain application process data, in each filtering process. In each filtering process, the gain applied to the filter application process data is acquired based on a target frequency characteristic of the image filtering process.