Image Processing Device Block-Based Resolution Conversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing devices require large memory capacity for decoding and resolution conversion of image data, leading to increased memory access time and restricted resolution options, while methods that reduce memory usage often introduce block noise, degrading image quality.
Innovation Solution
An image processing device that decodes and performs resolution conversion using a smaller memory capacity by cutting out and processing image data in a block-by-block manner, with horizontal and vertical resolution conversion sections, and employing mirroring processes to prevent noise at image edges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the entire decoded image data is temporarily stored in memory for resolution conversion, then resolution conversion can be performed on the decoded image data, but the memory capacity required becomes large and memory access time increases
Solution Approach 1:
The patent divides the image data into blocks (specifically 8x8 pixel blocks) and processes them individually through the resolution conversion process. Instead of storing the entire decoded image data in memory, only the current block being processed is held in the memory buffer, significantly reducing the required memory capacity while still enabling resolution conversion for the complete image.
Solution Approach 2:
The patent performs resolution conversion on blocks as they are decoded and become available, rather than waiting to process the entire image. This preliminary processing of blocks in the order they are generated allows the system to maintain small memory usage while achieving complete resolution conversion of the image data.
2Quantity of substance
If block-based resolution conversion is performed to reduce memory usage, then memory capacity requirement is reduced, but block noise is generated at block borders degrading image quality
Solution Approach 1:
The patent applies different processing characteristics to different regions of the image data. Specifically, when a block is positioned at the border of the image (having neighboring blocks in certain directions), special processing is applied to reduce block noise. The resolution conversion filter coefficients are adjusted based on whether the block is at the image border or in the interior, allowing localized optimization of image quality.
Solution Approach 2:
The patent uses feedback from the previously processed block to influence the processing of the current block. When a block has neighbors in both horizontal and vertical directions, the system uses information from adjacent blocks to refine the resolution conversion, reducing discontinuities and block noise at block borders through feedback-based filtering adjustments.
Data Source
AI summary
An image processing device and method decodes encoded image data, and converts a resolution of the decoded image data. The encoded image data is formed by encoding image data partitioned into blocks, each block including a plurality of pixels aligned in a matrix shape. The image processing device includes a decoding section that decodes and cuts out a portion of the decoded image data corresponding to a preset cut-out area and outputs the portion of the decoded image data on a block-by-block basis, a block buffer section that temporarily stores the portion of the decoded image data output, a horizontal resolution conversion section that converts the resolution of the decoded image data in the horizontal direction, a line buffer section that temporarily stores the horizontal-resolution-converted image data, and a vertical resolution conversion section that converts the resolution of the horizontal-resolution-converted image data in the vertical direction.