AI Upscaling Decoder Segmentation for Bitrate and Quality Trade-offs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for encoding and decoding high-resolution and high-definition images requires a codec that can efficiently handle the increased throughput of information, but existing methods struggle to improve encoding and decoding efficiency effectively.
Innovation Solution
A decoding apparatus and method that includes a processor to divide AI encoding data into image data and AI data, perform first decoding on the image data, and transmit the reconstructed image and AI data to an external apparatus, utilizing a deep neural network (DNN) for AI up-scaling, with options for transmission through HDMI or DisplayPort, and including information for HDR, color gamut, and resolution details.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If AI down-scaling followed by first encoding is used to compress images, then bitrate is reduced and transmission efficiency is improved, but image quality and resolution are degraded
Solution Approach 1:
The patent segments the image processing workflow into two distinct stages: first encoding the down-scaled image, then performing AI up-scaling on the decoded image. This segmentation allows each stage to be optimized independently - the encoding stage focuses on compression efficiency while the AI up-scaling stage focuses on quality restoration, resolving the contradiction between bitrate reduction and image quality preservation.
Solution Approach 2:
The patent introduces an AI up-scaling model as an intermediary component between the decoder and the final output. This intermediary takes the low-resolution decoded image and transforms it into a high-quality full-resolution image, acting as a bridge that recovers the quality loss incurred during compression without requiring the original high-bitrate stream.
2Manufacturing precision
If high-resolution images are encoded and transmitted, then image quality is maintained, but throughput and transmission bandwidth requirements increase
Solution Approach 1:
The patent inverts the conventional approach by first encoding a down-scaled version of the image and then applying AI up-scaling after decoding. Instead of transmitting full-resolution images and compressing them, the system transmits compressed low-resolution images and expands them using AI, achieving quality preservation with reduced throughput requirements.
3Manufacturing precision
If AI up-scaling is performed after decoding, then full-resolution image quality is restored, but additional processing time and computational resources are required
Solution Approach 1:
The patent performs AI up-scaling as a preliminary action during the decoding process rather than as a post-processing step. By integrating the up-scaling model into the decoding pipeline and using the decoded image as direct input, the system eliminates additional processing delays and enables real-time full-resolution output without significant time overhead.
Data Source
AI summary
Provided is a decoding apparatus including: a communication interface configured to receive AI encoding data generated as a result of artificial intelligence (AI) down-scaling and first encoding of an original image; a processor configured to divide the AI encoding data into image data and AI data; and an input/output (I/O) device, wherein the processor is further configured to: obtain a second image by performing first decoding on a first image obtained by performing AI down-scaling on the original image, based on the image data; and control the I/O device to transmit the second image and the AI data to an external apparatus. In some embodiments, the external apparatus performs an AI upscaling of the second image using the AI data, and displays the resulting third image.