♿ Detecting Vulnerable Pedestrian ♿

👉 BGSU opens a novel pedestrian dataset for vulnerable people

😎Review https://bit.ly/3JjVmu2
😎Paper arxiv.org/pdf/2212.06218.pdf
😎Data github.com/devvansh1997/BGVP

🧠 SERENA: LLM for Mental Health Support 🧠

👉Interactive #AI (in "#chatgpt" style) designed for mental health counseling

😎Review https://bit.ly/3wtbW37
😎Paper arxiv.org/pdf/2301.09412.pdf
😎Project https://serena.chat/

🐕 MAV3D: #3D Video from Text 🐕

👉#META unveils a novel #AI for generating #3D dynamic videos from text

😎Review https://bit.ly/3XN0zin
😎Paper arxiv.org/pdf/2301.11280.pdf
😎Project make-a-video3d.github.io

🔥CutLER: Unsupervised Segmentation 🔥

👉Novel paper by #META on detection & instance segmentation without human annotations

😎Review https://bit.ly/3DlFiUG
😎Paper arxiv.org/pdf/2301.11320.pdf
😎Code github.com/facebookresearch/CutLER
😎Project people.eecs.berkeley.edu/~xdwang/projects/CutLER

😍 CLIP/GPT3-driven Affective Faces 😍

👉Columbia unveils a neural framework for facial expressions retrieval given the context of the speaker

😎Review https://bit.ly/3HERna0
😎Paper arxiv.org/pdf/2301.10939.pdf
😎Project realtalk.cs.columbia.edu
😎Code github.com/scottgeng00/realtalk

🐦 Physics-inspired Computer Vision 🐦

👉UCLA unveils PhyCV, the first Physics-inspired Computer Vision Library

😎Review https://bit.ly/3HEWozI
😎Code github.com/JalaliLabUCLA/phycv
😎Project photonics.ucla.edu/2022/05/12/jalali-lab-open-sources-phycv-a-physics-inspired-computer-vision-library/

🎷Audio-Visual Semantic Segmentation🎷

👉A novel problem in #AI: pixel-level segmentation of objects that produce sound in the image frame

😎Review https://bit.ly/3wFY6dw
😎Paper arxiv.org/pdf/2301.13190.pdf
😎Project opennlplab.github.io/AVSBench
😎Code github.com/OpenNLPLab/AVSBench

🚛 Text-driven Video Neural Editing 🚛

👉A novel text-guided video editing with both appearance/shape

😎Review https://bit.ly/3YcfMJO
😎Paper arxiv.org/pdf/2301.13173.pdf
😎Project text-video-edit.github.io/

⭐ Mono-STAR: Unified Track/3D ⭐

👉Real-time 3D unified framework for semantic fusion, tracking, non-rigid deformation, and topological changes

😎Review https://bit.ly/3Dxvxmx
😎Paper arxiv.org/pdf/2301.13244.pdf
😎Project github.com/changhaonan/Mono-STAR-demo

🛋️🛋️ 100% Accurated #3D Labeling 🛋️🛋️

👉#Amazon unveils a novel tool for fine-grained 3D part labeling. Up to 100% accuracy! Paper only😢

😎Review https://bit.ly/3kYpQHQ
😎Paper https://arxiv.org/pdf/2301.10460.pdf

💧FLOW360: 360° Neural Optical Flow💧

👉 The first perceptually realistic 360° video benchmark dataset + SLOF method for OF tracking

😎Review https://bit.ly/3wMZZoX
😎Paper arxiv.org/pdf/2301.11880.pdf
😎Project https://siamlof.github.io

🐓DREAMIX:General Diffusive Video Editor🐓

👉#Google unveils the first diffusion-based method able to perform text-based motion/appearance editing of general videos

😎Review https://bit.ly/3I3Hq6B
😎Paper arxiv.org/pdf/2302.01329.pdf
😎Project dreamix-video-editing.github.io/

🧩 Text-Guided #3D Texturing 🧩

👉 Text-Guided HQ textures via iterative diffusion-based process

😎Review https://bit.ly/3ldC6Ez
😎Project texturepaper.github.io/TEXTurePaper
😎Code github.com/TEXTurePaper/TEXTurePaper
😎Paper texturepaper.github.io/TEXTurePaper/static/paper.pdf

🦚 MOSE: coMplex video Object SEgmentation 🦚

👉Novel Dataset for VOS is out! SOTA method on DAVIS is only 59.4% on MOSE

😎Review https://bit.ly/40yzSzW
😎Paper arxiv.org/pdf/2302.01872.pdf
😎Project henghuiding.github.io/MOSE/
😎Code github.com/henghuiding/MOSE-api

🌘 Gen-1: next-gen Generative #AI 🌘

👉#Runway unveils Gen-1: the next step forward for Generative AI. Registration available for beta -> hurry up!

😎Review https://bit.ly/3YqQYh8
😎Paper arxiv.org/pdf/2302.03011.pdf
😎Project https://research.runwayml.com/gen1

🗿DirectMHP: Multi-Head Pose Estimation🗿

👉Novel E2E multi-person head pose estimation (MPHPE) under full-range angles

😎Review https://bit.ly/3HJubXg
😎Paper arxiv.org/pdf/2302.01110.pdf
😎Code github.com/hnuzhy/DirectMHP

🧱 LEGO-Net: Objects in Rooms 🧱

👉Transformer-based iterative method for rearrangement of objects in messy rooms

😎Review https://bit.ly/3HR0fs6
😎Paper arxiv.org/pdf/2301.09629.pdf
😎Project ivl.cs.brown.edu/#/projects/lego-net

🎃 In-N-Out: 3D-aware OOD video editing 🎃

👉Novel 3D-aware video editing able to manipulate OOD objects (e.g. heavy makeup, accessories)

😎Review https://bit.ly/3jN0CMu
😎Paper arxiv.org/pdf/2302.04871.pdf
😎Project https://in-n-out-3d.github.io

🥸 MEGANE: Generative Morphable Eyeglass 🥸

👉#META unveils the most advanced #3D compositional morphable AI for eyeglasses (HD geometry/photometric interaction)

😎Review https://bit.ly/3jOWifu
😎Paper arxiv.org/pdf/2302.04868.pdf
😎Project junxuan-li.github.io/megane

💘 3D-aware Blending with NeRF 💘

👉Novel 3D-aware blending method via generative NeRFs

😎Review https://bit.ly/3lBEJA2
😎Paper arxiv.org/pdf/2302.06608.pdf
😎Project blandocs.github.io/blendnerf
😎Code github.com/naver-ai/BlendNeRF