20220418-20220501.md

成果推荐及讨论

• Haggai Maron

  • Sign and Basis Invariant Networks: this paper studies how to process Laplacian-based graph positional encodings in an invariant way. Very nice theory and strong experimental results!
  • paper: Sign and Basis Invariant Networks

• AK

  • Real-Time Neural Character Rendering with Pose-Guided Multiplane Images

    • paper: https://arxiv.org/abs/2204.11820
    • project page: https://ken-ouyang.github.io/cmpi/index.html

  • StyleGAN-Human: A Data-Centric Odyssey of Human Generation

    • project page: https://stylegan-human.github.io
    • github: https://github.com/stylegan-human/StyleGAN-Human

  • Deep Equilibrium Optical Flow Estimation

    • paper: https://arxiv.org/abs/2204.08442
    • github: https://github.com/locuslab/deq-flow

  • Photorealistic Monocular 3D Reconstruction of Humans Wearing Clothing

    • paper: https://arxiv.org/abs/2204.08906
    • project page: https://phorhum.github.io

  • What's in your hands? 3D Reconstruction of Generic Objects in Hands

    • paper: https://arxiv.org/abs/2204.07153
    • project page: https://judyye.github.io/ihoi/

  • NeuMips: Neural Mixture of Planar Experts for View Synthesis

    • paper: https://arxiv.org/abs/2204.13696
    • project page: https://zhihao-lin.github.io/neurmips/

• Timo Bolkart

  • Given a single image, EMOCA reconstructs a 3D face with facial expression detail that convey the emotional state of the input. (CVPR 2022)
    • Code & model: https://github.com/radekd91/emoca
    • paper: https://ps.is.mpg.de/uploads_file/attachment/attachment/686/EMOCA__CVPR22.pdf
    • Video: https://youtu.be/zjMLB2-dVGw

• Arash Vahdat

  • Critically-Damped Langevin Diffusion augments the input data space with auxiliary velocity variables, leading to smoother and simpler diffusion models.

    • Project page: https://nv-tlabs.github.io/CLD-SGM/
    • ICLR page: https://iclr.cc/virtual/2022/poster/6687

  • Denoising diffusion GAN uses a multimodal complex GAN denoising distribution for accelerating sampling from diffusion models.

    • Project page: https://nvlabs.github.io/denoising-diffusion-gan/
    • ICLR page: https://iclr.cc/virtual/2022/poster/7183

• Andraw Davison

  • The superpower of neural fields is scale-free representation of correlated high-dimensional information. In iLabel a small MLP jointly stores shape, colour, and many-channel open-set semantics; correlation is why dense segmentation can be trained with just a few real-time clicks.
    • Project page: https://edgarsucar.github.io/ilabel/

• Michael Black

  • Given the outside surface of the human body, can we peer inside and infer the bones? Many methods predict a “skeleton” that is not realistic. With OSSO CVPR2022, for the first time, we learn to predict a detailed skeleton from external observations.

    • Project: https://osso.is.tue.mpg.de
    • Code: https://github.com/MarilynKeller/OSSO
    • Paper: https://download.is.tue.mpg.de/osso/OSSO.pdf
    • Data: see project page
    • Video: https://youtube.com/watch?v=1wIiPVNDXY0

  • BEV (#CVPR2022) computes all the 3D people in an image in one shot, placing them all appropriately in depth. The key novelty is an imaginary 2D “Bird’s-Eye-View” (BEV) representation that reasons about the body centers in depth.

    • Code: https://github.com/Arthur151/ROMP
    • Project: https://arthur151.github.io/BEV/BEV.html
    • RH dataset: https://github.com/Arthur151/Relative_Human
    • paper: https://arxiv.org/abs/2112.08274

• Gerard Pons-Moll

  • BEHAVE: The first dataset with reliable 3D Human + Objects paired with images / depth.
  • Project: https://virtualhumans.mpi-inf.mpg.de/behave/

• Chiaraplizzari

  • Egocentric action recognition has just met event-based cameras.
  • paper: E^2(GO)MOTION: Motion Augmented Event Stream for Egocentric Action Recognition (CVPR 2022)

课程和报告分享

Hao Su Lab

• Generalizable Policy Learning in the Physical World[https://ai-workshops.github.io/generalizable-policy-learning-in-the-physical-world/]
• Time: Apr 29th @ 8:00 AM PDT (UTC -7)

Matthias Niessner

• A great overview about hot topics & open challenges in neural rendering, neural radiance fields, and many more!
• paper: https://arxiv.org/pdf/2111.05849
• report: https://youtu.be/Ry6hyn7ITtw

意见分享及讨论

工具分享

• UE5 Group Paint Tool
• This is useful to quickly create/edit PolyGroups on areas of a triangulated mesh, which can then be locally edited with the PolyEd tool. Also includes a brief demo of our new UV Editor!

招聘

• Davide Scaramuzza
  • They have multiple openings for Phd students and Postdocs in machine learning for computer vision and vision-based robot navigation. Job descriptions and how to apply: https://rpg.ifi.uzh.ch/positions.html