Converting 2D to 3D holographic art
How I converted my 2D Drawings to Holographic \ Lightfield 3D using A.I.
drive.google.com/open?id=1AZDBhydR7dqZdU2k0UxnQySRQTh5vW1C
Download 3D images from this video (zipped) or at the end of this post.
Download 3D images from this video (zipped) or at the end of this post.
Setup & Background
The Looking Glass display uses greyscale depth images to create a 3D effect that can be viewed from a wide angle. You can see the depth in the video as I rotate the screen, its amazing in person, it's in real stereo 3D.
Is it a hologram or lightfield? More info here https://lookingglassfactory.com/
about Google AI https://mannequin-depth.github.io/
Abstract
...a method for predicting dense depth in scenarios where both a monocular camera and people in the scene are freely moving. Existing methods for recovering depth for dynamic, non-rigid objects from monocular video impose strong assumptions on the objects’ motion and may only recover sparse depth. In this paper, we take a data-driven approach and learn human depth priors from a new source of data: thousands of Internet videos of people imitating mannequins, i.e., freezing in diverse, natural poses, while a hand-held camera tours the scene...
I started looking for depth tools... I found this Google AI conversion process but it was a bit too manual for me, no fun.
The Looking Glass display uses greyscale depth images to create a 3D effect that can be viewed from a wide angle. You can see the depth in the video as I rotate the screen, its amazing in person, it's in real stereo 3D.
Is it a hologram or lightfield? More info here https://lookingglassfactory.com/
about Google AI https://mannequin-depth.github.io/
Abstract
...a method for predicting dense depth in scenarios where both a monocular camera and people in the scene are freely moving. Existing methods for recovering depth for dynamic, non-rigid objects from monocular video impose strong assumptions on the objects’ motion and may only recover sparse depth. In this paper, we take a data-driven approach and learn human depth priors from a new source of data: thousands of Internet videos of people imitating mannequins, i.e., freezing in diverse, natural poses, while a hand-held camera tours the scene...
I started looking for depth tools... I found this Google AI conversion process but it was a bit too manual for me, no fun.
Depthmap generation from Google by Ai
I ran into this video on how to use Google AI to create a depth map using Google collab.
This is very usable but really cumbersome and slow, but I super appreciate the video and info!
It uses a virtual computer on a website to create the maps using the above tools.
I converted a group of images and got some very fun results.
Thanks to Stereo Maker!
I found this software that also uses AI to estimate and create depth maps, it's designed for panoramic images but works well on standard sizes.
Big thanks to Kaj Toet! I'm looking forward to see future updates.
How to create a depth map
1. Download and setup Holokit Deep360 from the link above
2. Add an image in the select input file
1. Download and setup Holokit Deep360 from the link above
2. Add an image in the select input file
3. Use these settings to start, the "write depth" option is where you get your depth map.
4. Set the output folder
5. Click START to create a depth map, it will do its best to evaluate the image and estimate real depth. It will create a two panel image and depth.
6. Open the dual image and depth map in an image editor.
7. Place a copy of the original color image at the top and the new depth map below. You want to replace the color because the app distorts it.
8. You should have something like this:
4. Set the output folder
5. Click START to create a depth map, it will do its best to evaluate the image and estimate real depth. It will create a two panel image and depth.
6. Open the dual image and depth map in an image editor.
7. Place a copy of the original color image at the top and the new depth map below. You want to replace the color because the app distorts it.
8. You should have something like this:
9. open The Looking Glass Depth media player, drag the image in and view!
