Image Deformation Device Using Segmented Memory Transfer
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-resolution image output devices, such as printers, face challenges in correcting complex image deformations due to the large memory requirements for storing positional correction data for each pixel, leading to increased costs and reduced precision in image deformation processes.
Innovation Solution
An image deformation device utilizing a primary memory for storing positional correction data and a smaller reference memory, with a transfer unit that sequentially transfers data in a predetermined order, allowing for precise pixel correction without the need for extensive SRAM capacity in the image processing LSI.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If positional correction amount data for each pixel is stored in SRAM of the image processing LSI, then precise positional correction can be achieved, but the memory capacity is increased and the LSI size is enlarged
Solution Approach 1:
The patent divides the large-capacity memory into two segments: primary memory (external to LSI) for storing all positional correction data, and reference memory (small SRAM within LSI) for storing only the current line's correction data. This segmentation allows precise correction while keeping LSI size manageable.
Solution Approach 2:
The patent introduces a temporal dimension to the memory architecture by sequentially transferring data from primary memory to reference memory line-by-line. This allows the system to use a small SRAM effectively by loading data in the time dimension rather than requiring all data to be simultaneously available.
2Measurement precision
If positional correction data for high-resolution images is stored, then correction precision is improved, but the memory capacity requirement increases significantly
Solution Approach 1:
The patent extracts only the necessary portion of positional correction data (current line data) from the complete dataset and loads it into the reference memory. This extraction approach allows high-resolution correction precision while using minimal SRAM capacity within the LSI.
Solution Approach 2:
The patent performs preliminary action by pre-calculating and storing all positional correction data in the primary memory before the actual image processing begins. This allows the LSI to work with pre-prepared correction data, reducing the real-time memory requirements during processing.
3Device complexity
If simple deformation correction is used, then the processing is simpler and faster, but the actual complex deformation of output images cannot be completely eliminated
Solution Approach 1:
The patent creates a copy of the ideal output image positions and compares it with the actual output positions to generate precise correction data. This copying approach allows the system to account for complex real-world deformations while maintaining a relatively simple processing architecture.
Data Source
AI summary
Disclosed is an image deformation device, including: a primary memory in which a positional correction amount data for each pixel, is stored; a reference memory having a smaller storing area than the primary memory; a transfer unit configured to transfer the positional correction amount data stored in the primary memory to the reference memory separately in multiple times; and a positional correction unit configured to carry out a positional correction process, wherein the transfer unit transfers the positional correction amount data according to the predetermined order and a progress of the positional correction process in the positional correction unit so that the positional correction amount data for one pixel has been already stored in the reference memory when the positional correction unit corrects the position of the one pixel.