🌨️Multi NeRF-supervised Depth-Pose🌨️

👉NeRF-supervised disentanglement of depth/camera pose from large-scale clips

😎Review https://bit.ly/3gS4Z7e
😎Paper arxiv.org/pdf/2210.07181.pdf
😎Project oasisyang.github.io/self-mpinerf

🟧 PlanT: Driving Plan w/ Transformer 🟧

👉PlanT: planning via transformer. Imitation learning with compact object-level representation

😎Review https://bit.ly/3Ni2MxN
😎Project www.katrinrenz.de/plant/
😎Paper arxiv.org/pdf/2210.14222.pdf
😎Code github.com/autonomousvision/plant

🔥Network -> C++ Converter. Pure Fire!🔥

👉Fully open source #Python framework to convert Neural Nets into C++. Lightning-fast inference💥

😎Review https://bit.ly/3gSyqWI
😎Code github.com/facebookincubator/AITemplate

💥TAVA for photo-realistic Avatars💥

👉#META unveils a novel approach for Template-free Animatable Volumetric Avatars. #Metaverse ready.

😎Review https://bit.ly/3DPDEvm
😎Paper arxiv.org/pdf/2206.08929.pdf
😎Project www.liruilong.cn/projects/tava/
😎Code github.com/facebookresearch/tava

🐀 DANNCE: animal keypoints tracking 🐀

👉DANNCE: anatomical #3D tracking across species and behaviors

😎Review https://bit.ly/3Nrx0P3
😎Code https://github.com/spoonsso/dannce

🥻Multi-Garment Virtual Try On (VTON)🥻

👉URJC unveils ULNeFs, a novel approach to efficiently solve mix-and-match VTO for multiple garment layers. SUIT UP!

😎Review https://bit.ly/3sVVAOC
😎Project mslab.es/projects/ULNeF

⚠️ #AI art is making creatives obsolete ⚠️

1. Scientist: "AI, let's make creatives obsolete"
2. AI: "Ok, let's do it! We have a deal"

👉 The AI 👆

😎More https://bit.ly/3DBWXa2

🩳 STC: Transformers MOT 🩳

👉A novel Multi-Object Tracking based on Transformers with Dense Representations

😎Review https://bit.ly/3WzQsNB
😎Paper arxiv.org/pdf/2210.13570.pdf
😎Code github.com/amitgalor18/STC_Tracker

🔥#AIwithPapers: we are 4,700!🔥

😎 Me, literally, every day 👆

😈 Share -> https://bit.ly/3fA5geu
😈 Invite friends -> https://t.iss.one/AI_DeepLearning

🔏First Depth/anti-FacRec Cam 🔏

👉First working prototype that enables passive depth estimation while inhibiting face identification

😎Code (coming)
😎Project https://zaidtas.github.io/privacymask
😎Paper zaidtas.github.io/privacymask/pdf/privacy_preserving_depth_estimation.pdf

😈 Synthetic Expression-Wrinkles 😈

👉#Microsoft unveils a novel approach that produces realistic wrinkles across humans

😎Review https://bit.ly/3zWZLOd
😎Paper arxiv.org/pdf/2210.03529.pdf
😎Project microsoft.github.io/DynamicWrinkles

🐛FactorMatte Video re-Composition🐛

👉FactorMatte: alternative formulation of video matting problem for re-composition

😎Review https://bit.ly/3NX6hdg
😎Paper arxiv.org/pdf/2211.02145.pdf
😎Project https://factormatte.github.io/
😎Code github.com/jacklyngu/FactorMatte

🍉 Unifying Flow, Stereo & Depth 🍉

👉A unified neural model for motion and #3D perception tasks

😎Review https://bit.ly/3AbdnFk
😎Paper arxiv.org/pdf/2211.05783.pdf
😎Project haofeixu.github.io/unimatch/
😎Code github.com/autonomousvision/unimatch

🔫 Real-Time Auto Neural Mapping 🔫

👉Robot can densely annotate a scene via fully-autonomous experimental interactions

😎Review https://bit.ly/3GfwQbY
😎Paper arxiv.org/pdf/2210.17325.pdf
😎Project ihaughton.github.io/RobE

🏔️ InfiniteNature-Zero: ∞ Video 🏔️

👉Open-source generative #AI for unbounded flythrough videos of natural scenes

😎Review https://bit.ly/3hJYELa
😎Paper arxiv.org/pdf/2207.11148.pdf
😎Project infinite-nature-zero.github.io
😎Code https://github.com/google-research/google-research/tree/master/infinite_nature_zero

🧶Open-Source Neural Inpainting Suite🧶

👉Lama Cleaner: a free and open-source inpainting tool powered by SOTA #AI model

😎Review https://bit.ly/3Ex2aSq
😎Project github.com/Sanster/lama-cleaner

🎃 eDiff-I: Generative AI by #Nvidia 🎃

👉eDiff-I: generative AI for text-to-image with instant style transfer & "paint-with-words"

😎Review https://bit.ly/3UEESzk
😎Project deepimagination.cc/eDiffi
😎Paper arxiv.org/pdf/2211.01324.pdf

♾ The Universal Image Segmentation ♾

👉OneFormer is the universal framework that unifies: semantic, instance & panoptic 🤯

😎Review https://bit.ly/3g4OAfD
😎Project praeclarumjj3.github.io/oneformer
😎Paper arxiv.org/pdf/2211.06220.pdf
😎Code github.com/SHI-Labs/OneFormer

🔥 ROVIS: Robust Online VIS 🔥

👉ROVIS is a novel method for robust online VIS that exhibits impressive accuracy on long clips

😎Review https://bit.ly/3UMugyv
😎Project zitongzhan.github.io/rovis_page/
😎Paper arxiv.org/pdf/2211.09108.pdf