Using AI to remove objects from photos and video
Researchers have figured out a way to use AI to fill in gaps in images, essentially automating the ability to seamlessly erase objects from photos or video. You have to watch the demonstration to truly appreciate how amazing it is. Their detailed mumbo-jumbo explanation:
We present a new flow-based video completion algorithm. Previous flow completion methods are often unable to retain the sharpness of motion boundaries. Our method first extracts and completes motion edges, and then uses them to guide piecewise-smooth flow completion with sharp edges. Existing methods propagate colors among local flow connections between adjacent frames. However, not all missing regions in a video can be reached in this way because the motion boundaries form impenetrable barriers. Our method alleviates this problem by introducing non-local flow connections to temporally distant frames, enabling propagating video content over motion boundaries. We validate our approach on the DAVIS dataset. Both visual and quantitative results show that our method compares favorably against the state-of-the-art algorithms.
I never have any idea what these guys are talking about, but the results speak for themselves. I'm just glad I finally have a way to remove my ex-girlfriend from all of my old pictures. Haha just kidding, I've never had a girlfriend. What...does love feel like?
Keep going for the full video demonstration. It's so impressive it seems fake.
(via Waxy)
Can you change the color of the dots with your mind?
This illusion won second place in the 2019 Best Illusion of the Year contest and it has sufficiently made me question reality: In this illusion, either downward...
This guy built an automated pizza assembly line out of Legos
To celebrate 5 years on YouTube, The Brick Wall assembled this Lego pizza-assembler using Legos. I wasn't expecting much, but the end result actually looks impressively like...
Creating 3D models of people from 2D images
Continuing humanity's race towards potential deepfake hell, researchers have developed a way of creating 3D models from 2D images using neural networks. The full title of the...
Using internet photos to create 3D reconstructions
Researchers have developed a system for taking photo collections from the internet and turning them into 3D reconstructions. They call it NeRF-W, which stands for Neural Radiance...
Researchers develop battery-free Game Boy that can run forever
Researchers at Northwestern University and the Delft University of Technology have developed a Game Boy that uses solar power and the energy generated from mashing buttons to...
Disney develops high resolution neural face-swapping
The folks at Disney Research Studios have moved one step closer to replacing human actors by developing this algorithm for high resolution face-swapping in images and videos....