Your Daily AI Research tl;dr | 2022-06-08

Welcome to your official daily AI research tl;dr (and news) intended for AI professionals and enthusiasts.

In this newsletter, I share the most exciting papers I find on a daily basis, along with a short summary to help you quickly seize if the paper is worth investigating. I will also take this opportunity to share daily interesting news in the field. I hope you enjoy the format of this newsletter, and I would gladly take any feedback you have in the comments to improve it.

Now, let's get started with this iteration!

1?? Learning to Generate Artistic Character Line Drawing

As you saw on the thumbnail of the newsletter, this new technique creates artistic character line drawings from input images, and the results are quite impressive (see more results in the paper)!

This is "just" another image-to-image translation problem where simply transforming an image into another "style", in this case, a drawing style. They used a newly designed GAN-based architecture to achieve that; GANs are still alive!! They also released a new dataset with "pairs of freehand character line drawings as well as corresponding character images/photos, where these line drawings with diverse styles are manually drawn by skilled artists."

You can download the pre-trained model, use their code and even have access to their dataset. Everything is shared publicly. What a cool way to do research!

Link to the paper: https://arxiv.org/pdf/2206.02998.pdf

Code: https://github.com/cnyvfang/P2LDGAN

2?? Layered Depth Refinement with Mask Guidance

Depth maps... Maybe the most important feature for 3D rendering or simply 2D images with cool 3D effects we've all seen on social media. Having a high-quality depth map is quite complex when you only have access to the images themselves - even though some techniques exist to do that (single image depth estimation (SIDE)), they are far from perfect. This model utilizes a generic mask to refine the depth prediction of these SIDE models, transforming the early results into high-quality depth maps.

They propose a self-supervised learning scheme that uses arbitrary masks and RGB-D datasets as datasets with depth annotations are pretty limited.

This is good news for background removal approaches or other

See more results: https://sooyekim.github.io/MaskDepth/

Link to the paper: https://arxiv.org/pdf/2206.03048.pdf

?? Researchers from China have developed a fully automated method to create pig clones

"Creating clones using human labor often results in damaged cells. Robotic pig cloning has a much higher success rate. The technology could help significantly increase China's pig population and help the country become self-sufficient in pork production."

What do you think of this?... ??

Read more in this great article by Rupendra Brahambhatt: https://interestingengineering.com/cloning-pigs-only-using-ai

And we are already at the end of this iteration! Please subscribe and share it with your techy friends if you've enjoyed it. Once again, let me know how to improve this format as this is something I have wanted to do for quite some time and haven't figured out the best way to do so. I hope you liked the decisions here, and I would be glad to hear from you to make it even better with time.

Thank you for reading, a fellow AI enthusiast and researcher.