Multi-Frame Super Resolution Precision Control During Gameplay
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-frame super resolution (MFSR) techniques in video gaming systems face challenges such as high computational load leading to reduced frame rates, inconsistent resolution, and increased processing overhead, resulting in a less stable and responsive gaming experience.
Innovation Solution
A method and system that dynamically adjusts the precision of weights and activations in a trained neural network based on processing unit usage, using multiple neural networks with varying precisions to maintain frame rate and image quality during high computational loads.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If multi-frame super resolution (MFSR) is performed using high precision weights and activations, then image quality and resolution are improved, but computational load increases and frame rate decreases
Solution Approach 1:
The system dynamically adjusts the precision of neural network weights and activations based on real-time processing unit usage monitoring. When computational load is high, precision is reduced to maintain frame rate; when load is low, precision is increased to improve image quality. This dynamic adaptation resolves the contradiction between maintaining high image quality and achieving high frame rate.
Solution Approach 2:
The invention changes the precision parameter of neural network weights and activations based on system conditions. By varying precision levels (e.g., from high precision to lower precision representations), the system can trade off between computational speed and image quality, effectively resolving the contradiction between frame rate and image quality.
2Productivity
If dynamic resolution reduction is used to maintain frame rate during high load, then frame rate is maintained, but resolution consistency deteriorates and image detail is lost
Solution Approach 1:
Instead of changing resolution, the system changes the precision parameter of neural network computations. This allows frame rate to be maintained during high load while keeping resolution consistent, as the same resolution is displayed but with varying computational precision behind the scenes.
3Measurement precision
If high precision neural network computations are performed, then image accuracy is improved, but processing time increases and responsiveness decreases
Solution Approach 1:
The system dynamically adjusts computation precision based on processing unit usage and timing requirements. By monitoring system state in real-time, the system can switch between high precision (when time permits) and lower precision (when responsiveness is critical), resolving the contradiction between image accuracy and processing speed.
Data Source
Figure 1A~1B
Figure 2~3
Figure 4~5
AI summary
A computer-implemented method of generating a multi-frame super resolution, MFSR, graphics output for a video gaming system during gameplay, comprising: using a trained artificial neural network, ANN, comprising a plurality of weights and activations to perform multi-frame super resolution, MFSR, based on input graphics data from a game deployed on a video gaming system, to generate a MFSR graphics output; monitoring usage of a processing unit of the video gaming system during gameplay; and varying a precision of the weights and/or activations used in the performing MFSR, based on the monitored usage of the processing unit. A corresponding video gaming system and computer program product is also provided.