🤯 Point·E: the "DALL·E" of 3D pts cloud 🤯

👉OpenAI opens Point·E: the "evolution" of DALL·E in the #3D point cloud domain

😎Review https://bit.ly/3I87PR9
😎Code github.com/openai/point-e
😎Paper arxiv.org/pdf/2212.08751.pdf

🥑InstantAvatar: Avatar in 1 min.🥑

👉Avatars from monocular video in ONE minute (130x faster than SOTA)

😎Review https://bit.ly/3PNw1tJ
😎Paper arxiv.org/pdf/2212.10550.pdf
😎Project tijiang13.github.io/InstantAvatar

🐬 DeepLSD: Neural Lines Detector 🐬

👉 DeepLSD: novel neural line detector in the wild. Paper, Code, and Models available

😎Review https://bit.ly/3WqONK2
😎Paper arxiv.org/pdf/2212.07766.pdf
😎Code github.com/cvg/deeplsd

❄️ ClimateNeRF: Extreme Climate NeRF ❄️

👉NeRF-based realistic climate change simulations for disaster forecasting

😎Review https://bit.ly/3vhqdz6
😎Paper arxiv.org/pdf/2211.13226.pdf
😎Project https://climatenerf.github.io

☃️ Neural Smooth Video Composition ☃️

👉The novel SOTA in neural video composition. From a few frames to infinite!

😎Review https://bit.ly/3hRyaIe
😎Code github.com/genforce/StyleSV
😎Paper arxiv.org/pdf/2212.07413.pdf
😎Project genforce.github.io/StyleSV/

🎆 One-Shot Video Generation 🎆

👉Tune-A-Video: One-Shot Video Generation with no training pain!

😎Review https://bit.ly/3hX47il
😎Paper arxiv.org/pdf/2212.11565.pdf
😎Project https://tuneavideo.github.io/
😎Code github.com/showlab/Tune-A-Video

🍯 Interpretable Shape Programs 🍯

👉GeoCode: #3D shape synthesis via intuitively editable parameters

😎Review https://bit.ly/3PYgc3o
😎Paper arxiv.org/pdf/2212.11715.pdf
😎Code github.com/threedle/GeoCode
😎Project threedle.github.io/GeoCode

🎇 Real-Time NeRF+SLAM 🎇

👉Novel geometric/photometric 3D mapping pipeline powered by NeRF+SLAM

😎Review https://bit.ly/3G6xEOO
😎Paper arxiv.org/pdf/2210.13641.pdf
😎Code github.com/ToniRV/NeRF-SLAM

💡MetaPortrait: ID-preserving talking head💡

👉USTC unveils a novel ID-preserving talking head generation framework

😎Review https://bit.ly/3jyVWsW
😎Paper arxiv.org/pdf/2212.08062.pdf
😎Project https://meta-portrait.github.io
😎Code github.com/Meta-Portrait/MetaPortrait

⚛️ Event-based vehicle tracking ⚛️

👉Novel event-based vehicle detection/tracking framework. Dataset included!

😎Review https://bit.ly/3vzxWJ7
😎Paper
arxiv.org/ftp/arxiv/papers/2212/2212.14289.pdf
😎Dataset
sar-lab.net/event-based-vehicle-detection-and-tracking-dataset

⚠️ ALERT: #PyTorch COMPROMISED ⚠️

[EDIT: THE ALERT IS NOT CONFIRMED, IT COULD BE JUST A BUG]

👉PyTorch-nightly packages installed via pip compromised with malicious binary 🦠

😎More https://bit.ly/3WEm5p1

🍏 MUSE: T2I Generative Transformers 🍏

👉Google unveils the new SOTA (faster and better) for text-to-image generation

😎Review https://bit.ly/3CeEu3B
😎Paper arxiv.org/pdf/2301.00704.pdf
😎Project muse-model.github.io

🔢 #AI Advanced Math Reasoning 🔢

👉#Meta unveils HTPS: SOTA transformer-based theorem prover.

😎Review https://bit.ly/3CoUsrP
😎Paper arxiv.org/pdf/2205.11491.pdf
😎Blog ai.facebook.com/blog/ai-math-theorem-proving

🔥 YOLOv5 Segmentation in OpenCV 🔥

👉#OpenCV recently added the support for YOLOv5 instance segmentation

😎Review https://bit.ly/3idVNeb
😎Code github.com/ultralytics/yolov5
😎Tutorial learnopencv.com/yolov5-instance-segmentation

🚀 Nvidia + ByteDance CV-CUDA 🚀

👉CV-CUDA: open-source, GPU-accelerated lib for CV developed jointly by NVIDIA / ByteDance

😎Review https://bit.ly/3GHSaqu
😎Code github.com/CVCUDA/CV-CUDA

🔥 Source Code of Box2mask is out! 🔥

👉BoxInstSeg is a toolbox for SOTA box-supervised instance segmentation

😎Review https://bit.ly/3QryBGf
😎Code github.com/LiWentomng/BoxInstSeg

🗣️ MemFace: Generative Talking Face 🗣️

👉#Microsoft (+SJTU) unveils MemFace: the new SOTA in talking faces generation

😎Review https://bit.ly/3k8TjhZ
😎Paper arxiv.org/pdf/2212.05005v2.pdf
😎Project memoryface.github.io/

🪛 AgileAvatar: HQ stylized 3D #Avatar 🪛

👉#ByteDance unveils a novel self-supervised framework for #3D avatars

😎Review https://bit.ly/3kaOAw6
😎Project ssangx.github.io/projects/agileavatar
😎Paper ssangx.github.io/pubs/2022-SIGGRAPHAsia-AgileAvatar.pdf

🔥 #Nvidia GC ViT Large -> updated! 🔥

👉The new SOTA (with large margin) across image classification, detection & semantic segmentation!

😎Review https://bit.ly/3W4ChyF
😎Code github.com/NVlabs/GCViT
😎Paper arxiv.org/pdf/2206.09959.pdf