Bi-Direction Sub-Filter Design for Lower-Power Image Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Performing image filtering operations on entire kernels of image frames is computationally expensive and increases power consumption in digital cameras.
Innovation Solution
Divide kernels into sub-kernels and perform filtering operations on each sub-kernel, using previously calculated values for overlapping sub-kernels to reduce computational load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If image filtering operation is performed on entire kernel, then filtered value is obtained, but computational cost and power consumption increase
Solution Approach 1:
The patent divides the kernel into multiple sub-kernels (e.g., 4 sub-kernels for a 2x2 kernel) and performs filtering operations on each sub-kernel separately. The filtered values from sub-kernels are then combined to produce the final filtered value for the entire kernel. This segmentation reduces the computational load per operation while maintaining accuracy.
Solution Approach 2:
Instead of performing filtering on the entire kernel at once, the patent applies filtering to partial sub-kernels only. By processing smaller portions of the kernel, the processor consumes less power for each filtering operation while still achieving the same overall filtering effect through combination of partial results.
2Reliability
If image filtering operation is performed on entire kernel, then filtered value is obtained, but number of computational operations increases
Solution Approach 1:
The kernel is segmented into multiple sub-kernels, and filtering operations are performed on each sub-kernel independently. This segmentation reduces the number of computational operations required compared to processing the entire kernel at once, as each sub-kernel can be processed with fewer operations.
Solution Approach 2:
The patent performs filtering operations on partial sub-kernels rather than the entire kernel. This partial action approach reduces computational complexity and the number of operations needed while maintaining the same filtering accuracy through proper combination of results.
3Productivity
If filtering operation is performed on overlapping sub-kernels, then computational load is reduced, but redundant calculations occur
Solution Approach 1:
The patent performs filtering operations on sub-kernels in a systematic order, and when a sub-kernel is found to be identical to a previously processed sub-kernel (due to overlapping kernels), the algorithm skips redundant calculations by using the pre-computed filtered value. This preliminary action approach avoids unnecessary energy consumption.
Solution Approach 2:
The algorithm uses previously computed filtered values from overlapping sub-kernels to serve the current processing needs. When a sub-kernel overlaps with previously processed areas, the system reuses existing filtered values rather than performing duplicate calculations, thereby reducing energy consumption.
Data Source
AI summary
A device may determine a plurality of filtered values for a plurality of sub-kernels of a kernel of image data, the filtered values corresponding to performing an image filtering operation on each of the plurality of sub-kernels, including using one or more filtered values for one or more sub-kernels of a neighboring kernel as a first one or more filtered values for a first one or more sub-kernels of the kernel and performing the image filtering operation on each of a second one or more sub-kernels of the kernel to generate a second one or more filtered values for the second one or more sub-kernels of the kernel. The device may determine, using the plurality of filtered values, a filtered value for the kernel that corresponds to performing the image filtering operation on the kernel.