@topher_batty@mastodon.acm.org avatar

topher_batty

@topher_batty@mastodon.acm.org

Associate Professor in CS at University of Waterloo doing computer graphics and computational physics. I also manage https://physicsbasedanimation.com and maintain a list of graphics researchers on Mastodon: https://christopherbatty.github.io/Mastodon-Graphics-Academics/

This profile is from a federated server and may be incomplete. Browse more on the original instance.

jbigham, to random
@jbigham@hci.social avatar

politics is a lot more interesting than we give it credit for … kudos to Governor Beshear, and kudos to the buckeyes

https://www.nytimes.com/live/2023/11/07/us/election-day-2023

topher_batty,
@topher_batty@mastodon.acm.org avatar

@jbigham well, now we won't have to worry about Joe Manchin anymore.

topher_batty, to random
@topher_batty@mastodon.acm.org avatar

The code for our "Reach for the Spheres" SDF reconstruction paper is now online: https://github.com/sgsellan/reach-for-the-spheres/

topher_batty, to random
@topher_batty@mastodon.acm.org avatar

For computer graphics academics and researchers, I threw this list of folks together. Apologies if I've forgotten anyone or included anyone who prefers to be forgotten! Let me know if there are any errors or corrections. Hoping this will help folks migrating from the birdsite. 🙂
https://christopherbatty.github.io/Mastodon-Graphics-Academics/

topher_batty,
@topher_batty@mastodon.acm.org avatar

@aeva Thanks! I put it together mostly by hand because whatever version of mastodon my server is on doesn't seem to exports lists. Ah well.

topher_batty,
@topher_batty@mastodon.acm.org avatar

@aeva I have fixed a couple more trailing @'s and other silly things. Thanks again!

scheidegger, to random
@scheidegger@mastodon.social avatar

I was expecting stupid, and this is worse. It's stupid and anodyne. Fuck this guy.

topher_batty,
@topher_batty@mastodon.acm.org avatar

@scheidegger @jennybryan literal LOL. I stopped reading long before that point I guess.

topher_batty,
@topher_batty@mastodon.acm.org avatar

@scheidegger @jennybryan no doubt. But yeah, that's the guy that founded Guelph.

eniko, to random
@eniko@peoplemaking.games avatar

Why did ant man quantumania get such bad reviews? It was a perfectly fine MCU movie

topher_batty,
@topher_batty@mastodon.acm.org avatar

@eniko I can't tell if this is a compliment to Ant Man, or a devastating takedown of the MCU movies.😆

Onsequitur, to random
@Onsequitur@genart.social avatar

@topher_batty @silviasellan do you know of any documentation on the derivation of triangle normal/area gradients.

Specifically for deriving gradient constraints: area conservation and normal orientation.

I've looked at the problem a few times and realize that there is a tedious method of expanding all variables then chugging away, then there is the differential geometry/matix calculus method. I figure I could brush up on the latter. Thanks!

topher_batty,
@topher_batty@mastodon.acm.org avatar

@Onsequitur I did triangle area gradients via index notation in the supplemental material for my 2012 viscous sheets paper. Tedious. I'm sure Keenan Crane has notes somewhere (for SIGGRAPH or a CMU class) that do it in a clever geometric way. IIRC it comes out to the vector perpendicular to the edge opposite the vertex, but don't quote me.

eniko, (edited ) to random
@eniko@peoplemaking.games avatar

Periodic reminder that if you're a leftist and you don't wear a mask indoors in public; no you ain't

topher_batty,
@topher_batty@mastodon.acm.org avatar

@eniko I am not sold on the idea that snarky leftist gatekeeping will win more people to that cause, but I guess worth a try.

scheidegger, to random
@scheidegger@mastodon.social avatar

From a former academic to other academics: y'all really need to unlearn the habit of starting a non-academic request with a backhanded compliment.

The rest of the world doesn't operate on that currency!

topher_batty,
@topher_batty@mastodon.acm.org avatar

@scheidegger can you give an example? It surprises me that anyone would insult you and then ask a favor. Is it really a thing?

topher_batty, to random
@topher_batty@mastodon.acm.org avatar

I'm very excited to share our new SIGGRAPH Asia 2023 paper on improved reconstruction of explicit surface meshes from signed distance field data: "Reach for the Spheres: Tangency-Aware Surface Reconstruction of SDFs".

https://odedstein.com/projects/reach-for-the-spheres/

topher_batty,
@topher_batty@mastodon.acm.org avatar

The key idea of our "Reach for the Spheres" algorithm is to interpret each SDF sample point and its associated distance value as representing a sphere (or circle in 2D). Then, a valid reconstructed surface that satisfies the SDF data must never enter any sphere, the sign of each sphere dictates which side of the surface it lies on, and, critically, the surface must touch each sphere (at least) once.

For example, compare here a Marching Squares reconstruction vs. a valid surface.

topher_batty,
@topher_batty@mastodon.acm.org avatar

By seeking to satisfy these conditions, our algorithm can provide superior reconstructions (e.g., vs. Marching Cubes, Neural Dual Contouring) on incredibly low resolution SDF grids, such as this 10^3 grid example.

This was a fun collaboration with Silvia Sellán (Toronto) and Oded Stein (USC).

By the way, first author, U of T PhD student Silvia Sellán is on the job market this year, so be sure to check out her three other brand new papers, and also hire her!

topher_batty,
@topher_batty@mastodon.acm.org avatar

@nervous_jesse Thanks! Yeah, as long as you are doing simple linear or polynomial interpolation to (artificially) upsample the SDF for meshing, you won't get as good results, since the naive upsampling discards/damages the useful SDF properties. As for topology, yeah, that's something we are exploring next. :-)

topher_batty,
@topher_batty@mastodon.acm.org avatar

@aeva Certainly slower than something simple like Marching Cubes or (non-Neural) Dual Contouring, since it is based on gradually flowing an initial mesh towards a satisfying state and remeshing en route. Speed is something we hope can be improved in future work.
As for voxel density / layout, it really only assumes a collection of input SDF sample points, which can be laid out in space any way you like -- no explicit dependence on a grid or octree structure.

topher_batty,
@topher_batty@mastodon.acm.org avatar

@aeva I doubt, but what's your application? 2D or 3D?

topher_batty,
@topher_batty@mastodon.acm.org avatar

@narain Probably true in some cases, yes. In the 'closest point method' literature they call those 'closest point surfaces' (albeit also discarding the sign) which lets them work with open surfaces too. And I saw an ML paper or two advocating (reinventing) the concept as well. :-) Pricier to store of course.

topher_batty,
@topher_batty@mastodon.acm.org avatar

@narain I dug up this 2021 computer vision paper that uses what they call "gradient-SDFs" (https://arxiv.org/abs/2111.13652), but essentially the same idea appears in this CPM paper from 2008 https://www.sciencedirect.com/science/article/pii/S002199910700441X
(And probably one could find earlier instances if you really went spelunking in the literature. ) Definitely seems useful!

topher_batty,
@topher_batty@mastodon.acm.org avatar

@nervous_jesse @narain the swept volume example we show is basically a whole bunch of unions, so CSG is doable. I suspect offsets are also straightforward, if you're just uniformly inflating. Other operations I'm not sure, but worth exploring.

jbigham, to random
@jbigham@hci.social avatar

the ML research communities have some problems, but i do like how they try stuff with reviewing -- e.g., we submitted a rebuttal for our EMNLP paper, and now we have a whole week to have a conversation with the anonymous reviewers about it. we see their reactions in real time and can clarify things that aren't clear. seems to address a big problem w/ rebuttals. of course, this could go badly, but there's a lot of incentive to remain civil given the chairs know who people are.

topher_batty,
@topher_batty@mastodon.acm.org avatar

@jbigham SIGGRAPH rebuttal phase in the early 2000s was basically such a back and forth forum chat, including uploading new data, text, video clips, but it was incredibly stressful because it turned into SIGGRAPH: Lightning Round! Reviews would say "the method can't do X" or "why didn't you show Y" and you'd spend the whole week sleepless and trying to produce such results in a mad panic. Not recommended. :-)

topher_batty, to random
@topher_batty@mastodon.acm.org avatar

For those attending , be sure to get up early for the "Pushing the Boundaries" paper session on Wed at 9am. Ryusuke Sugimoto will present our work on a Monte Carlo raytracing-like algorithm for solving Laplace/Poisson problems called the Walk on Boundary method. https://rsugimoto.net/WoBforBVPsProject/ Then, you can start converting all your favorite rendering techniques into PDE solvers!

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • anitta
  • thenastyranch
  • rosin
  • GTA5RPClips
  • osvaldo12
  • love
  • Youngstown
  • slotface
  • khanakhh
  • everett
  • kavyap
  • mdbf
  • DreamBathrooms
  • ngwrru68w68
  • megavids
  • magazineikmin
  • InstantRegret
  • normalnudes
  • tacticalgear
  • cubers
  • ethstaker
  • modclub
  • cisconetworking
  • Durango
  • provamag3
  • tester
  • Leos
  • JUstTest
  • All magazines
    •