AI Image Scaling With Filter Set Selection for Bandwidth Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge of transmitting high-definition images is hindered by limited network resources and the disparity between increasing video capacity and stagnant network bandwidth, necessitating improved image compression and restoration techniques.
Innovation Solution
An electronic apparatus and server system utilize AI models to downscale and upscale image data, employing a filter set selection process to minimize differences between original and restored images, leveraging Convolutional Neural Networks (CNNs) for efficient compression and restoration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If image quality is improved through compression and restoration, then video capacity increases, but network bandwidth cannot keep up with the increase
Solution Approach 1:
The server performs AI-based upscaling in advance during the encoding process, pre-restoring the image quality before transmission. This preliminary action allows the network to transmit compressed data while the receiving end benefits from pre-processed high-quality images, resolving the bandwidth limitation without sacrificing image quality
Solution Approach 2:
The patent replaces traditional mechanical compression algorithms with AI-based deep learning models for image restoration. This substitution enables more efficient compression ratios and better quality preservation, allowing higher video capacity within the same network bandwidth constraints
2Manufacturing precision
If AI upscaling models with multiple filter sets are used, then image restoration quality improves, but processing complexity increases
Solution Approach 1:
The patent segments the filter sets into different categories (first filter sets for natural images, second filter sets for synthetic images) and selects appropriate filters based on image type. This segmentation reduces processing complexity by avoiding unnecessary filter applications while maintaining high restoration quality through targeted AI model selection
Solution Approach 2:
The system dynamically selects which AI upscaling model to apply based on the source image characteristics (natural vs. synthetic). This dynamic adaptation simplifies processing by choosing the most suitable model rather than processing all images through all possible filters, reducing computational overhead while maintaining optimal quality
Data Source
AI summary
The disclosure relates to an artificial intelligence (AI) system that uses a machine learning algorithm and an application thereof. A method for controlling an electronic apparatus according to the disclosure includes receiving image data and information associated with a filter set that is applied to an artificial intelligence model for upscaling the image data from an external server; decoding the image data; upscaling the decoded image data using a first artificial intelligence model that is obtained based on the information associated with the filter set; and providing the upscaled image data for output.