Julien Valentin retweeted
With #COVID19 many research internship will be cancelled... so why not participate in a Google Summer of Code #GSoC instead, be exposed to cutting edge research, and make some extra $$?
#TensorFlowGraphics have several projects posted here: summerofcode.withgoogle.com/…
Julien Valentin retweeted
Vote for state of the art baselines we will implement in TensorFlow Graphics during the Google Summer of Code!
forms.gle/ETHdcTVZVYZ3tUZXA
Julien Valentin retweeted
If you are at #NeurIPS2019 , don't miss our paper on sparse multi-view (1 or more) object reconstruction. **Wed Dec 11th 10:45 AM -- 12:45 PM @ East Exhibition Hall B + C #96** Joint work with: @davrempe @JPCValentin @_sofien_ Leo Guibas. @NeurIPSConf
Julien Valentin retweeted
ViewAL: active learning with viewpoint entropy! youtu.be/tAGdx2j-X_g
Only 7% training data, achieves 95% network performance!
Simple idea:
If network makes inconsistent predictions across views, we need more training data for these cases :)
#DeepLearning-now-needs-less-data
Julien Valentin retweeted
Find the #SIGGRAPH2019 course notes on Differentiable Graphics with TensorFlow 2.0 at julienvalentin.github.io/SIG…
@JPCValentin @_sofien_ @random_forests @shancarter @DynamicWebPaige @zzznah
@martin_wicke
Julien Valentin retweeted
Our Deep View Synthesis CVPR 2019 paper is now online augmentedperception.github.i… , with examples of Light Field Video youtu.be/UD7I9gXIVsc?t=13 and an interactive multi-plane image (MPI) viewer to try the view interpolation technique for yourself augmentedperception.github.i…
Julien Valentin retweeted
How can we estimate the 6D pose and size of novel object instances that we have never seen before? Come find how we do category-level 6D pose and size estimation at our #CVPR2019 oral today (~15:00, Promenade Ballroom, Poster #181). Code/data available! geometry.stanford.edu/projec…
Julien Valentin retweeted
On Thursday we are going to introduce TF Graphics, a new library that brings 3D and graphics functionalities to TensorFlow! The presentation will take place in the 'Cutting Edge TensorFlow'session of #io19. Tune in the livestream tinyurl.com/y3alvh6p
@JPCValentin
Julien Valentin retweeted
Some of our AR lighting technology will debut at Google I/O next week; research paper on the technique to appear at CVPR 2019 arxiv.org/abs/1904.01175 teapawt.org/wmafx/status/112374297…
Julien Valentin retweeted
Our new Siggraph'19 work: Deferred Neural Rendering: Image Synthesis using Neural Textures.
youtu.be/z-pVip6WeyY
We learn features that directly live on a 3D mesh, which allows us to exercise full control of photo-realistic re-synthesis :)
#deeplearning #siggraph2019 #DL #AI
Julien Valentin retweeted
Very excited about our new investment round @synthesiaIO !
The startup behind that deep-fake David Beckham video just raised $3M tcrn.ch/2XMT8bL via @techcrunch
Julien Valentin retweeted
It is not too late! Submit your best work and spread the word!!!
Third Workshop on Computer Vision for AR/VR (CVPR19)
Deadline for Submission: May 1, 2019
Magic Leap has kindly provided an ML1 Creator Edition device as a prize for the best spotlight!
mixedreality.cs.cornell.edu/…









