Hannah and I are working on some personal projects this month! Here are my goals (part 1!):
Project 1: web app for tracking sneaky non-vegetarian foods:
A friend and I started a facebook group years ago with information on (animal) rennet, a substance found in a lot of cheeses that isn’t actually vegetarian (animals must be killed to obtain this substance). The websites dedicated to rennet-free cheeses and places to eat (in NYC) are hardly ever updated and seem a bit old-school. Another similar substance, Isinglass, found in many beers will also be tracked.
I want to create a modern web application using something like flask or node.js / mongodb / twitter bootstrap and host it from heavyimage.com to allow anyone to contribute new discoveries that are rennet free. This project appeals to my interest in keeping my web chops up to date with modern tech, as well as my interest in providing accessible, up to date resources for vegetarians trying to stay honest.
If time allows, it’d be a good experience to build a RESTful API so other apps can access the data I’ve collected.
Project 2: Facetracking stuff:
I attended Dan Shiffman’s ITP Camp 2013 lecture on recent advances in face tracking software. I had three project ideas:
- A faceoff app: inspired by John Woo’s 1996 masterpiece “Face / Off”, every person tracked by the software would get their faces swapped with the person next to them, in real time, at the press of a button (probably the “/” key)
- A morphing app: I’ve actually always wanted to implement the code to perform morphing between two faces!
- Real time face replacement: if I work up to it, using open frameworks ala https://vimeo.com/29279198 except better, stronger and faster….I’m just interested in implimenting the tech and seeing if I can do better.
These three projects will also allow me to sharpen my vim skills as processing and I believe oF allow the use of external editors / compiler toolchains / command line workflows now.
Project 3: Nuke tetris:
Everyone’s seen pong implemented in the nuke nodegraph. I want to build tetris!
Project 4: Photogrammetry R+D:
I’ve become really interested in photogrammetry (also called Structure From Motion or sfm), the process whereby still images can be converted back into 3d objects / environments using the power of math. My friend Dan showed me 123d Catch, a free application from Autodesk that does this process for you in the cloud. After shooting about 20-40 pictures on your iPhone, you can send the images to the cloud where they’re processed by Autodesk’s proprietary system and a textured model is returned to you. I’m turned off by Autodesk’s black box approach and their limitations, namely downrezing your images to speed their ascent into the cloud, a max of 40 pictures, and output that is just a textured model. I was curious what the state of FOSS alternatives to this were, so I did some research.
By leveraging free, open source alternatives, I want explore some implications of photogrammetry software:
- Nuke integration. I’d like to build a workflow to start with as many images as you like and end up with geo in nuke with projector cameras already all set up, (since the positions of the cameras are calculated during the process (and then just discarded by 123D)) projecting the photos onto the geo as a starting point for building a digital set without the need for expensive LIDAR scanners.
- Accelerated point tracking. Another possible application of learning about structure-from-motion software is borrowing their accelerated GPU-based feature tracking algorithms which could possibly be jerry rigged to provide accelerated point tracking for 2d / 3d tracking in compositing software.
- Building models of things from movies. I’m interested in the idea feeding a sequence from a film into photogrammmetry software to build a 3d model of a set. If there are angles that aren’t visible, it’d be fun to try to recreate it based on exploiting symmetry or reflections and produce a gallery (3d printed? WebGL?) of the set as a physical space you’re allowed to explore without the constraints of the original camera work.
Stay tuned, I’ve already made some promising progress!
Project 5: Abstract Geo Renders:
My website is littered with my art: abstract 3d geometry rendered experimentally to produce interesting results. I want to continue these processes and try to create some new renders. Possible new aspects will be:
- Produce 2d/3d prints at high quality.
- I’d love to try to emulate the look of lee bul’s work for some of the forms
- Experiment more with HDR lighting to instantly produce more realistic lighting for the scene.
- In high school, I produced other abstract renders using Bryce, which no longer runs on my computer. However, I’ve gotten it working using an emulator, sheepshaver, so I’d love to re-render my old scenes at acceptable modern resolutions.
- Document, possibly in tutorial form, my process for producing these shapes / renders.
- Build an Amazon EC2 image that comes set up with all the rendering software so I can quickly render my images on the cloud.