Image Enhancement Pipeline for Cropped Video Size Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image-enhancement processing technologies face difficulties in accurately identifying and processing images with original sizes due to overscan cropping, leading to challenges in setting up circuits and requiring modifications in neural network training data when dealing with cropped images.
Innovation Solution
An image-enhancement processing system that employs multiple cropping processes to identify image sizes before cropping, allowing for consistent setup and processing by employing an enhancement preprocessor, postprocessor, and composer through firmware and software collaboration, with a pseudo timing generator to manage frame delays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Shape
If the display system crops the screen image during video playback to meet output aspect ratio requirements, then the output video can have the correct aspect ratio, but the image enhancement processing cannot accurately identify the original image size for high-quality processing
Solution Approach 1:
The patent applies preliminary action by identifying and recording the original image size information before the cropping operation occurs. The system captures the original resolution data (e.g., 3840×2160) prior to any aspect ratio adjustment, ensuring that enhancement processing can later reference this original dimension information even after the image has been cropped to a different aspect ratio.
Solution Approach 2:
The patent introduces an intermediary mechanism in the form of a timing generator and firmware layer that manages the transition between original and cropped image dimensions. This intermediary layer maintains the relationship between original image size and cropped output, allowing enhancement processors to access original dimension information through controlled interfaces without directly processing the cropped image data.
2Adaptability or versatility
If the system processes cropped images for enhancement, then the processing can adapt to output requirements, but the circuit setup becomes complex and requires modifications to neural network training data
Solution Approach 1:
The patent extracts the dimension identification and timing management functions from the complex circuit setup into a dedicated firmware layer. By separating the concern of original image size identification from the enhancement processing circuitry, the system reduces circuit complexity while maintaining adaptability to different output formats through software-based timing control.
Solution Approach 2:
The patent creates a universal timing generator and firmware interface that can handle multiple output formats and aspect ratios through a single standardized mechanism. This multi-functional approach allows the same circuit setup to adapt to various output requirements without requiring format-specific hardware configurations or separate neural network training data for each scenario.
3Productivity
If the enhancement processing is performed on cropped images without proper timing management, then processing can proceed, but frame synchronization issues occur causing out-of-sync composition
Solution Approach 1:
The patent implements feedback mechanisms through the timing generator that continuously monitors and adjusts frame timing information throughout the enhancement processing pipeline. The timing generator provides real-time synchronization signals that ensure each processed frame is correctly timed and synchronized with the original video stream, preventing out-of-sync composition while maintaining processing throughput.
Solution Approach 2:
The patent applies preliminary timing setup and frame delay management before the enhancement processing occurs. By pre-configuring the timing generator with original frame rate and timing information, and by implementing frame delay buffers in advance, the system ensures that synchronization is maintained throughout processing without requiring complex real-time adjustments that could compromise reliability.
Data Source
AI summary
A method for image enhancement and an image-enhancement processing system are provided. In the method, an input video with an original resolution is obtained. The input video is frame-by-frame cropped into a selected video with an output resolution according to setting of an output video. The selected video is saved to a memory. On the other hand, enhancement preprocessing is performed on each frame of the selected video, so as to form multiple layers of downscaled images with different reduced proportions. One or more image-enhancement-processing processes are performed on the multiple layers of downscaled images. After that, enhancement post-processing is performed on the multiple layers of downscaled images, so that these downscaled images in each frame of the selected video are mapped to be mapping-layer images with the output resolution. Lastly, the mapping-layer images and the selected video are frame-by-frame composed, so as to generate the output video.