Director of XR Presence at Facebook Reality Labs
San Francisco, CA
Joined December 2017
- Tweets58
- Following222
- Followers666
- Likes1.1K
Sofien Bouaziz retweeted
Introducing HyperNeRF! By lifting NeRF into higher dimensions, we fix a big limitation of Nerfies: topological changes. HyperNeRF can handle topological variations such as facial expressions that Nerfies can't.
Website: hypernerf.github.io
arXiv: arxiv.org/abs/2106.13228
(1/N)
Sofien Bouaziz retweeted
Today we're releasing the code for Nerfies!
We've also included a Colab demo so that you can try out a basic version of our method right in your browser.
Project Page: nerfies.github.io
Github: github.com/google/nerfies
[1/n] 🧵👇
Sofien Bouaziz retweeted
Experience Augmented Reality like never before! With your smartphone! #depth #depthfrommotion #augmentedreality #smartphoneAR #AR 9to5google.com/2020/06/25/an…
Sofien Bouaziz retweeted
Depth sensing without a depth sensor, on hundreds of millions Android phones. The ARCore Depth API rolls out today → goo.gle/ARCore-Depth-API
Sofien Bouaziz retweeted
Join us at @cvpr2020! This year we organized a tutorial on how to leverage performance capture for perception tasks. We also talk about neural rendering and its applications! More info at seanfanello.it/2020/06/06/cv… #deeplearning #neuralrendering #volumetriccapture #computergraphics
Sofien Bouaziz retweeted
Check out our tutorial on Disentangled 3D Representations for Relightable Performance Capture of Humans ! #CVPR2020 augmentedperception.github.i…
From high quality depth sensing to neural rendering!
Sofien Bouaziz 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
Sofien Bouaziz 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
Sofien Bouaziz retweeted
New research leverages @TensorFlow Lite with MediaPipe to enable on-device, real-time hand tracking and gesture recognition, critical for applications such as #AR and on-device #signlanguage understanding. Check it out with the #opensource pipeline here → goo.gle/30hzVAD
Sofien Bouaziz retweeted
False color maps are used in diverse applications from medical imaging to #AR. Check out Turbo, a new colormap that capitalizes on the advantages of rainbow maps while solving shortcomings, such as false detail, banding, and color blindness ambiguity. goo.gle/2YZETof
Sofien Bouaziz retweeted
TensorFlow Graphics is out! This new library brings 3D and graphics functionalities to #TensorFlow!
@_sofien_ Cem Keskin @podlipensky @kiamada @AvneeshSud
#io19 #DeepLearning #GoogleARVR #GoogleAI
goo.gle/2DVFyLy
Sofien Bouaziz retweeted
Call for Posters/Extended Abstracts/Demos - 3rd Workshop on Computer Vision for AR/VR - June 17, 2019, Long Beach, CA mixedreality.cs.cornell.edu/…
Sofien Bouaziz retweeted
This is a fantastic resource. It was written before the deep revolution and therefore provides good context for it all. I spent most of the the first year of my PhD jumping from chapter to chapter of this book.
"Pattern Recognition and Machine Learning" by @ChrisBishopMSFT is now available as a free download. Download your copy today for an introduction to the fields of pattern recognition & machine learning: aka.ms/prml #ML #Insights
Sofien Bouaziz retweeted
"Everybody Dance Now" from Caroline Chan, Alyosha Efros and team transfers dance moves from one subject to another
The only way I'll ever dance well. Amazing work!!!
youtu.be/PCBTZh41Ris?t=2m14s
arxiv.org/abs/1808.07371
Sofien Bouaziz retweeted
code: github.com/NVIDIA/vid2vid
video: goo.gl/c6Em4R
Using Conditional GAN to generate HD resolution videos
"Video-to-Video Synthesis," Wang et al.: arxiv.org/abs/1808.06601
Impressive stuff from NVIDIA.
Sofien Bouaziz retweeted
If you want to know how probabilistic inference can be tied to optimal control, I just put up a new tutorial on control as inference: arxiv.org/abs/1805.00909
This expands on the control as inference lecture in my class: rll.berkeley.edu/deeprlcours…
Sofien Bouaziz retweeted
Pretty mind-blowing paper, 3D-printed representations of neural networks you can run inference on by shining light through... literal *light speed* inference times.
They only have classification networks so far, but I'd love to see pix2pix or a GAN soon.
science.sciencemag.org/conte…
Sofien Bouaziz retweeted
Learning Human Optical Flow. They use a 3D model of the human body and motion capture data to synthesise realistic flow fields. arxiv.org/pdf/1806.05666.pdf.
Sofien Bouaziz retweeted
Code for the Glow demo is now on github! Use the pre-trained model to create manipulation vectors from your own labelled dataset (eg: changing hair color to red). You can also run the demo locally, go make your own smiling videos!
github.com/openai/glow