Between things like camera shake and poor lighting, the videos you take with your phone, DSLR or mirrorless camera can end up blurry for any numbers of reasons. Worse yet, it’s difficult to sharpen a photo or video after the fact, and more often than not, the results don’t look great. However, with the help of AI, Adobe thinks it may have the solution for blurry photos and videos.
During the Sneaks portion of its Max 2020 conference, the company showed off an experimental feature called Sharp Shots that’s powered by its Sensei AI. It uses machine learning to deblur each frame of a video. The results can be a bit hard to appreciate in a compressed YouTube clip, but they mostly speak for themselves. You can see in the videos that Adobe shared that there’s a significant difference in image clarity. It’s most noticeable in the final example, with the facial features of the two friends in the clip much easier to make out in the AI-processed video.
One thing to keep in mind is we may not see Sharp Shots ever make its way to an Adobe product. It’s also worth pointing out we’ve seen other companies promise a lot with AI and then not deliver. In one of its most famous I/O demos, Google showed off a Photos feature it was working on that would allow you to remove objects from images. The company never ended up shipping that feature.
That said, AI is enabling some of the most compelling new features coming to Adobe’s products. In Photoshop, for example, AI will soon allow you to add more dramatic skies to your photos. That’s something that won’t change even if Adobe doesn’t end up shipping Sharp Shots.
Author: Igor Bonifacic, @igorbonifacic
Source: Engadget
