Video Frame Fusion with Selective Super-Resolution at High Zoom
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mobile devices suffer from low video resolution due to high scaling factors during video shooting, which existing technologies have not adequately addressed.
Innovation Solution
A video processing method that alternates exposure times for video frames, applies a super-resolution algorithm to one frame, and fuses frames with different weights to enhance resolution, using a first zoom mode for high scaling and a second mode for low scaling or dark scenes, and incorporates a logarithmic curve and look-up table processing for style enhancement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If super-resolution algorithm is applied to both first and second exposure frames, then video resolution is improved, but processing time and computational complexity increase
Solution Approach 1:
The patent applies super-resolution algorithm partially only to the first exposure frame rather than both frames. This partial application reduces processing time and computational complexity while still achieving improved video resolution through the fusion of the processed first frame and unprocessed second frame with optimized weights.
2Measurement precision
If fusion weight of first exposure frame is increased, then resolution enhancement effect is improved, but image quality and natural appearance deteriorate
Solution Approach 1:
The patent dynamically adjusts fusion weights based on scene luminance conditions. In dark scenes, the weight of the first exposure frame is reduced to preserve image quality and natural appearance. In bright scenes, the weight is increased to enhance resolution enhancement effect. This dynamic adaptation resolves the contradiction between resolution enhancement and image quality preservation.
3Illumination intensity
If exposure time of first frame is increased, then light collection is improved, but motion blur increases
Solution Approach 1:
The patent segments the video frames into two types: first exposure frames with longer exposure time for better light collection in dark scenes, and second exposure frames with shorter exposure time to minimize motion blur. By segmenting and processing these frames differently with optimized fusion weights, the patent simultaneously achieves good light collection and reduced motion blur effects.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
Embodiments of this application relate to the field of video shooting technologies, and provide a video processing method and apparatus, an electronic device, and a storage medium, which can reduce a problem of a low resolution due to a high scaling factor. The video processing method includes: in a first zoom mode or a second zoom mode, obtaining a video shot through a camera lens; in the first zoom mode, processing a video image of a first exposure frame by using a super-resolution algorithm, where processing by using the super-resolution algorithm is used for increasing a resolution; in the first zoom mode, fusing a video image of a first exposure frame processed by using the super-resolution algorithm and a video image of a second exposure frame not processed by using the super-resolution algorithm, to obtain a fused video, where the video image of a first exposure frame has a first weight; and in the second zoom mode, fusing the video image of a first exposure frame and the video image of a second exposure frame, to obtain a fused video, where the video image of a first exposure frame has a second weight, and the first weight is greater than the second weight.