Multi-Frame AI Super-Resolution for Camera Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing super-resolution techniques face challenges in obtaining high-resolution images, particularly in devices like smartphones, due to hardware limitations and the limitations of single image super-resolution methods.
Innovation Solution
A method using an artificial intelligence model to align multiple low-resolution images obtained from a camera and a reference image to generate a high-resolution image, incorporating training with actual camera data to improve the super-resolution technique.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If pixel shift method is used to obtain high-resolution image from multiple low-resolution images, then image resolution is improved, but device complexity increases due to requiring hardware components for camera movement
Solution Approach 1:
The patent replaces the mechanical pixel shift method with an artificial intelligence-based image processing system. Instead of physically moving the camera sensor to capture multiple images at different pixel positions, the system uses AI algorithms to analyze and reconstruct high-resolution images from standard camera captures, thereby eliminating the need for complex mechanical movement components.
Solution Approach 2:
The patent introduces an artificial intelligence model as an intermediary between the low-resolution input images and the final high-resolution output. This AI mediator processes the images through learned patterns and transformations, achieving super-resolution without requiring direct mechanical pixel shifting hardware.
2Device complexity
If single image super-resolution method is used, then device complexity is reduced, but image resolution improvement is limited
Solution Approach 1:
The patent employs multiple successive image captures and processes them continuously through the AI model. Instead of relying on a single image, the system captures a sequence of images and applies continuous processing to accumulate useful information, thereby achieving better resolution improvement while keeping hardware simple.
Solution Approach 2:
The patent merges multiple low-resolution images captured in succession into a single high-resolution image through AI processing. By combining information from multiple input images, the system achieves superior resolution enhancement compared to single-image methods, while maintaining hardware simplicity.
3Manufacturing precision
If multiple images are captured and aligned to improve resolution, then image quality is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary alignment and preprocessing of multiple captured images before applying the full AI super-resolution processing. By pre-aligning images and preparing them in advance, the system reduces the computational burden during the main processing stage, thereby minimizing processing time while maintaining high image quality.
Data Source
AI summary
An electronic device according to various embodiments of the disclosure may include: a camera module including a camera, a memory, and at least one processor electrically connected to the camera module and the memory, wherein the at least one processor may be configured to: successively obtain a plurality of first images through the camera module, align the plurality of first images, obtain a reference image based on the plurality of aligned first images, and obtain a second image using an artificial intelligence model based on the plurality of aligned first images and the reference image.