Image Processing Apparatus Fourier Transform Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing apparatuses, especially in portable devices, face challenges with high computational demands and time consumption due to the need for numerous Fourier transform operations when processing hologram images, which limits their performance in terms of size and power availability.
Innovation Solution
The implementation of an image processing apparatus that performs Fourier transform operations efficiently by using a layer-based algorithm, parallel processing with multiple cores, and a method of performing 2D Fourier transforms through sequential 1D Fourier transforms in row and column directions, along with pixel encoding and image quality correction, to reduce calculation time and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image processing apparatuses perform Fourier transformation operations to reproduce hologram images, then the quality and detail of hologram reproduction is improved, but the calculation time and power consumption increase significantly
Solution Approach 1:
The patent segments the 2D Fourier transform operation into two separate 1D Fourier transform operations - one for row directions and one for column directions. This segmentation reduces the computational complexity from O(N² log N) for direct 2D FFT to a more efficient calculation approach, thereby reducing power consumption while maintaining hologram image quality
Solution Approach 2:
The patent transforms the problem from a direct 2D Fourier transform approach to a sequential 1D transform approach by introducing a dimensional decomposition strategy. By processing rows and columns separately in sequence rather than simultaneously, the computational burden is reduced, enabling efficient hologram generation in resource-constrained portable devices
2Measurement precision
If image processing apparatuses perform multiple Fourier transform operations for each pixel location, then the accuracy of hologram reproduction is improved, but the processing time increases
Solution Approach 1:
The patent divides the complex 2D Fourier transform into sequential 1D transform operations for row and column directions. This segmentation allows the processing apparatus to perform multiple transform operations with high accuracy while reducing the time required for each operation, as 1D transforms are computationally less intensive than 2D transforms
Solution Approach 2:
The patent performs preliminary 1D Fourier transform operations on row data before performing the second 1D transform on column data. This preliminary action approach allows the system to prepare intermediate results that can be efficiently processed in the second transform stage, maintaining accuracy while optimizing processing time through staged computation
3Adaptability or versatility
If portable devices perform Fourier transformation operations for hologram generation, then the functionality and image quality are improved, but the device size and power requirements increase
Solution Approach 1:
The patent applies segmentation by breaking down the computationally intensive 2D Fourier transform into two manageable 1D transform operations. This reduces the peak power consumption and total energy required for hologram generation, making the functionality viable for portable devices with limited power availability while maintaining full hologram generation capability
Solution Approach 2:
The patent implements a dynamic processing approach where the system adaptively performs sequential 1D transforms rather than a static 2D transform. This dynamic computation strategy optimizes power usage by processing data in stages, allowing portable devices to generate holograms efficiently within their power constraints while preserving the adaptability and versatility of the hologram generation function
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An image processing apparatus includes a core configured to perform a fast Fourier transformation (FFT) operation on the image data, a memory configured to store data that is output by the core, and a controller configured to control the core to perform the FFT operation on the image data. The core is resettable based on an amount of the image data.