GAN Media Enhancement Distribution Cloud Device
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face challenges in enhancing media content to a desired quality, as existing methods either require high network bandwidth when enhancing on a cloud server or consume excessive processing resources and battery power when enhancing locally on a user computing device, without optimizing the distribution of Generative Adversarial Network (GAN) processing between these locations.
Innovation Solution
A method that determines the required degree of enhancement for media content and calculates the necessary processing time and buffer requirements, identifying optimal distribution between a cloud-hosted server and user computing device based on factors like network bandwidth, data plan, and cloud service costs, to efficiently enhance media content using a GAN, either on the cloud server, user device, or a combination of both.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If media content enhancement is performed on a cloud-hosted server, then enhancement quality is improved, but network bandwidth consumption increases
Solution Approach 1:
The patent divides the GAN enhancement process into two segments: a first GAN model runs on the cloud-hosted server to perform initial enhancement, while a second GAN model runs on the user computing device to perform final enhancement. This segmentation allows the system to balance cloud processing (for quality) with local processing (to reduce network bandwidth consumption), as the enhanced content only needs to be transmitted once rather than multiple times through the network.
Solution Approach 2:
The patent introduces a new dimension of processing distribution by deploying GAN models across two different locations (cloud and device) rather than concentrating all enhancement operations in one place. This dimensional change in processing architecture enables simultaneous optimization of enhancement quality and network bandwidth efficiency.
2Quantity of substance
If media content enhancement is performed locally on user computing device, then network bandwidth consumption is reduced, but processing resources and battery power are excessively consumed
Solution Approach 1:
The enhancement workload is segmented between cloud and device: the cloud-hosted server performs the computationally intensive first enhancement pass using its first GAN model, while the user device performs a lighter second enhancement pass using its second GAN model. This segmentation reduces the processing burden and energy consumption on the user device compared to performing all enhancement operations locally.
Solution Approach 2:
The cloud-hosted server acts as an intermediary by performing preliminary enhancement before transmitting content to the user device. This intermediary processing reduces the amount of work the user device must perform, thereby reducing its energy and processing resource consumption.
3Device complexity
If a single GAN model is used for enhancement, then system complexity is reduced, but enhancement effectiveness and quality are insufficient
Solution Approach 1:
The system uses two separate GAN models instead of one: a first GAN model on the cloud-hosted server for initial enhancement and a second GAN model on the user device for final enhancement. This segmentation of the enhancement function into multiple specialized models improves overall enhancement effectiveness and quality while keeping each individual model relatively simple.
Solution Approach 2:
The patent combines the capabilities of two GAN models operating at different locations (cloud and device) to achieve superior enhancement quality. By merging the processing power of both models in a distributed architecture, the system achieves better results than a single model could provide alone.
Data Source
AI summary
In an approach for optimizing the enhancement of media content using a Generative Adversarial Network (GAN), a processor identifies a degree of enhancement to be performed on the media content to attain the defined quality. A processor calculates a length of time required to enhance the media content to the degree of enhancement identified considering one or more factors. A processor identifies a buffer required to enhance the media content to the degree of enhancement identified, wherein the buffer prevents an interruption in the media content. A processor identifies a distribution to enhance the media content between a cloud hosted server and a user computing device. A processor enhances the media content according to the distribution identified.