Thursday 23 January 2025
Scientists have been working on a new technique to clear blurry videos, and it’s getting close to perfection. The team used a combination of sharpness detection and edge information to develop an algorithm that can remove blur from video footage.
The researchers created a model called SPEINet, which stands for Spatial-Temporal Edge Information Network. This model is designed to detect the sharpest frames in a video and then use that information to clear up the blurry parts. The team tested their algorithm on videos captured by GoPro cameras, which are known for producing high-quality footage.
The results were impressive. SPEINet was able to clear up blur from 80% of the test videos, making them look almost as good as they did when they were originally recorded. The team also tested other algorithms and found that their own model performed better than the others in most cases.
But how does it work? Well, the first step is to detect the sharpest frames in a video. This is done by analyzing the edge information in each frame, which helps to identify areas where there is clear detail. The team used a technique called edge detection, which involves looking for changes in brightness and color between adjacent pixels.
Once the sharpest frames have been identified, the algorithm uses that information to clear up the blurry parts of the video. This is done by interpolating missing data from the sharp frames and filling in gaps where there was no clear detail. The result is a video that looks like it was recorded with a high-quality camera.
The team’s algorithm has many potential applications. For example, it could be used to restore old videos that have become blurry over time. It could also be used to improve the quality of video footage captured by lower-end cameras or in low-light conditions.
The researchers plan to continue improving their algorithm and testing it on different types of videos. They hope to make it more efficient and effective, so that it can be used for a wide range of applications.
Cite this article: “Sharpening Blurry Videos with Advanced Algorithm”, The Science Archive, 2025.
Blurry Videos, Video Processing, Algorithm, Edge Detection, Interpolation, Sharpness Detection, Speinet, Gopro Cameras, Video Restoration, Image Enhancement
