This media is not supported in your browser
VIEW IN TELEGRAM
đ¸ Segment Anything in HQ đ¸
đHQ-SAM: SAM with the ability to accurately segment objects, maintaining promptable design, efficiency, zero-shot generalizability
đReview https://t.ly/GxX5B
đPaper arxiv.org/pdf/2306.01567.pdf
đModels github.com/SysCV/SAM-HQ
đHQ-SAM: SAM with the ability to accurately segment objects, maintaining promptable design, efficiency, zero-shot generalizability
đReview https://t.ly/GxX5B
đPaper arxiv.org/pdf/2306.01567.pdf
đModels github.com/SysCV/SAM-HQ
đĨ18đ4đ¤¯1đą1đ1
This media is not supported in your browser
VIEW IN TELEGRAM
đ Track Everything Everywhere đ
đ#Google unveils OmniMotion: full-length motion tracking for every pixel in every frame of video.
đReview https://t.ly/Krvw
đPaper arxiv.org/pdf/2306.05422.pdf
đProject omnimotion.github.io/
đDemo omnimotion.github.io/#interactive_demo
đCode github.com/qianqianwang68/omnimotion
đ#Google unveils OmniMotion: full-length motion tracking for every pixel in every frame of video.
đReview https://t.ly/Krvw
đPaper arxiv.org/pdf/2306.05422.pdf
đProject omnimotion.github.io/
đDemo omnimotion.github.io/#interactive_demo
đCode github.com/qianqianwang68/omnimotion
đĨ23â¤5đ¤¯3đ¤Š1đŠ1
This media is not supported in your browser
VIEW IN TELEGRAM
đī¸ Scene Five: Through Her Eyes đī¸
đ #3D scene reconstruction of what a person is observing using only the reflections of their eyes
đReview https://t.ly/uBO6
đPaper arxiv.org/pdf/2306.09348.pdf
đProject https://world-from-eyes.github.io/
đ #3D scene reconstruction of what a person is observing using only the reflections of their eyes
đReview https://t.ly/uBO6
đPaper arxiv.org/pdf/2306.09348.pdf
đProject https://world-from-eyes.github.io/
đ¤¯28đĨ12đŠ2đ¤Š1
This media is not supported in your browser
VIEW IN TELEGRAM
đ§ŋ NeRF-Supervised Deep Stereo đ§ŋ
đA novel pioneering pipeline for training deep stereo networks WITH NO ground-truth
đReview https://t.ly/c7j-
đProject nerfstereo.github.io/
đDataset https://amsacta.unibo.it/id/eprint/7218/
đCode github.com/fabiotosi92/NeRF-Supervised-Deep-Stereo
đPaper https://openaccess.thecvf.com/content/CVPR2023/papers/Tosi_NeRF-Supervised_Deep_Stereo_CVPR_2023_paper.pdf
đA novel pioneering pipeline for training deep stereo networks WITH NO ground-truth
đReview https://t.ly/c7j-
đProject nerfstereo.github.io/
đDataset https://amsacta.unibo.it/id/eprint/7218/
đCode github.com/fabiotosi92/NeRF-Supervised-Deep-Stereo
đPaper https://openaccess.thecvf.com/content/CVPR2023/papers/Tosi_NeRF-Supervised_Deep_Stereo_CVPR_2023_paper.pdf
đĨ°8đ¤Š3â¤1đ1đŠ1đ1
This media is not supported in your browser
VIEW IN TELEGRAM
đĢŖ Text-Guided Adversarial Makeup đĢŖ
đNovel facial privacy protection via adversarial latent codes. Makeup vs Face Recognition.
đReview https://t.ly/pBCP
đPaper arxiv.org/pdf/2306.10008.pdf
đCode github.com/fahadshamshad/Clip2Protect
đNovel facial privacy protection via adversarial latent codes. Makeup vs Face Recognition.
đReview https://t.ly/pBCP
đPaper arxiv.org/pdf/2306.10008.pdf
đCode github.com/fahadshamshad/Clip2Protect
â¤6đ1đĨ1đĨ°1đŠ1
Media is too big
VIEW IN TELEGRAM
đĻˇ Few-Shot Geometry-Aware Keypoints đĻˇ
đUBC (+Flawless AI) unveils the new SOTA in semantic keypoints localization. Suitable for faces, animals, cars, mouth, teeth & more
đReview https://t.ly/-0qN
đPaper arxiv.org/pdf/2303.17216.pdf
đProject xingzhehe.github.io/FewShot3DKP/
đUBC (+Flawless AI) unveils the new SOTA in semantic keypoints localization. Suitable for faces, animals, cars, mouth, teeth & more
đReview https://t.ly/-0qN
đPaper arxiv.org/pdf/2303.17216.pdf
đProject xingzhehe.github.io/FewShot3DKP/
đ¤¯10đ4â¤2âĄ2đ2đ¤Š2đĨ1
This media is not supported in your browser
VIEW IN TELEGRAM
đ Fooling Neural Forensic Classifiers đ
đAdversarial faces able to fool the forensic classifiers, while remaining undetectable by humans
đReview https://t.ly/33Cc
đPaper arxiv.org/pdf/2306.13091.pdf
đProject koushiksrivats.github.io/face_attribute_attack
đCode github.com/koushiksrivats/face_attribute_attack
đAdversarial faces able to fool the forensic classifiers, while remaining undetectable by humans
đReview https://t.ly/33Cc
đPaper arxiv.org/pdf/2306.13091.pdf
đProject koushiksrivats.github.io/face_attribute_attack
đCode github.com/koushiksrivats/face_attribute_attack
đĸ6â¤4đ2đą2đž2đ1đ¤¯1đ1
panohead_overview-min.gif
24.3 MB
đĨ PanoHead: 3D Full-Head Synthesis đĨ
đ#ByteDance (+UW-M) unveils PanoHead: 360âĻ view-consistent portraits from a single-view image
đReview https://t.ly/MrLNR
đPaper arxiv.org/pdf/2303.13071.pdf
đProject sizhean.github.io/panohead
đCode github.com/sizhean/panohead
đ#ByteDance (+UW-M) unveils PanoHead: 360âĻ view-consistent portraits from a single-view image
đReview https://t.ly/MrLNR
đPaper arxiv.org/pdf/2303.13071.pdf
đProject sizhean.github.io/panohead
đCode github.com/sizhean/panohead
đĨ7â¤4đ¤¯3đą1
AI with Papers - Artificial Intelligence & Deep Learning
đ Drag-GAN: user-friendly image-manipulation đ đ Manual deforming of (real and generated) images over pose, shape, expression and layout. đReview https://bit.ly/3BFyXlR đPaper arxiv.org/pdf/2305.10973.pdf đProject vcai.mpi-inf.mpg.de/projects/DragGANâĻ
Linkedin
#google #artificialintelligence #machinelearning #ml #ai #deeplearningâĻ | Alessandro Ferrari | 40 comments
đĨđĨ Source Code of Drag-GAN IS OUT! đĨđĨ
đManual deforming of (real and generated) images over pose, shape, expression and layout. Source Code just released a few hours ago đ
đđĸđ đĄđĨđĸđ đĄđđŦ:
â Max Planck + MIT + #Google AR/VR = đ¤¯
â Supervising handle points to moveâĻ
đManual deforming of (real and generated) images over pose, shape, expression and layout. Source Code just released a few hours ago đ
đđĸđ đĄđĨđĸđ đĄđđŦ:
â Max Planck + MIT + #Google AR/VR = đ¤¯
â Supervising handle points to moveâĻ
đĨ25đą6â¤3đĨ°1đ¤¯1
This media is not supported in your browser
VIEW IN TELEGRAM
đŽSAM-PT: Segment Anything+TrackingđŽ
đSAM-PT is the first method to utilize sparse point propagation for Video Object Segmentation (VOS).
đReview https://t.ly/QLMG
đPaper arxiv.org/pdf/2307.01197.pdf
đProject www.vis.xyz/pub/sam-pt/
đCode github.com/SysCV/sam-pt
đSAM-PT is the first method to utilize sparse point propagation for Video Object Segmentation (VOS).
đReview https://t.ly/QLMG
đPaper arxiv.org/pdf/2307.01197.pdf
đProject www.vis.xyz/pub/sam-pt/
đCode github.com/SysCV/sam-pt
đĨ14â¤7đ¤¯3đ1đą1
This media is not supported in your browser
VIEW IN TELEGRAM
đĒŠ DISCO: Human Dance Generation đĒŠ
đNTU (+ #Microsoft) unveils DISCO: a big step towards the Human Dance Generation.
đReview https://t.ly/cNGX
đPaper arxiv.org/pdf/2307.00040.pdf
đProject disco-dance.github.io/
đCode github.com/Wangt-CN/DisCo
đNTU (+ #Microsoft) unveils DISCO: a big step towards the Human Dance Generation.
đReview https://t.ly/cNGX
đPaper arxiv.org/pdf/2307.00040.pdf
đProject disco-dance.github.io/
đCode github.com/Wangt-CN/DisCo
đĨ13đĨ°4đ2âĄ1đ1đž1
This media is not supported in your browser
VIEW IN TELEGRAM
đŖī¸ STAR.: 3D-tracking w/ attention paradigm đŖī¸
đ#Mercedes STAR: e2e 3D object tracking that follows the tracking-by-attention paradigm
đReview https://t.ly/JoGj
đPaper arxiv.org/pdf/2306.17602.pdf
đProject simondoll.github.io/publications/star_track
đ#Mercedes STAR: e2e 3D object tracking that follows the tracking-by-attention paradigm
đReview https://t.ly/JoGj
đPaper arxiv.org/pdf/2306.17602.pdf
đProject simondoll.github.io/publications/star_track
đ14đĨ1đĨ°1đ1
This media is not supported in your browser
VIEW IN TELEGRAM
đĄ Text2Cinemagraphs: Cinemagraph from text đĄ
đCMU (+ #Snap) unveils a fully automated method for creating cinemagraphs from text descriptions
đReview https://t.ly/BwZs6
đPaper arxiv.org/pdf/2307.03190.pdf
đProject text2cinemagraph.github.io/website
đCode github.com/text2cinemagraph/text2cinemagraph
đCMU (+ #Snap) unveils a fully automated method for creating cinemagraphs from text descriptions
đReview https://t.ly/BwZs6
đPaper arxiv.org/pdf/2307.03190.pdf
đProject text2cinemagraph.github.io/website
đCode github.com/text2cinemagraph/text2cinemagraph
â¤12đ¤¯3đą1đ¤Š1
This media is not supported in your browser
VIEW IN TELEGRAM
đĨTest-Time Training on fire đĨ
đExtending the TTT to the streaming setting. Suitable for Panoptic, Instance & Colorization.
đReview https://t.ly/eZYA
đPaper arxiv.org/pdf/2307.05014.pdf
đProject https://video-ttt.github.io/
đCode github.com/renwang435/video-ttt-release
đExtending the TTT to the streaming setting. Suitable for Panoptic, Instance & Colorization.
đReview https://t.ly/eZYA
đPaper arxiv.org/pdf/2307.05014.pdf
đProject https://video-ttt.github.io/
đCode github.com/renwang435/video-ttt-release
đĨ10đ3âĄ1đ¤¯1
This media is not supported in your browser
VIEW IN TELEGRAM
đ Deepfake via casual self-scan đ
đTAU presents a novel approach to reenact an ID using only a casual self-scan
đReview https://t.ly/9T8Wi
đPaper arxiv.org/pdf/2307.06307.pdf
đProject arielazary.github.io/PGR
đTAU presents a novel approach to reenact an ID using only a casual self-scan
đReview https://t.ly/9T8Wi
đPaper arxiv.org/pdf/2307.06307.pdf
đProject arielazary.github.io/PGR
đ¤¯7đ6â¤5đĨ1đą1
This media is not supported in your browser
VIEW IN TELEGRAM
đĒ Extreme Human Pose Estimation đĒ
đRePoGen: novel synthetic data generator of extreme/realistic poses of humans
đReview https://t.ly/ecBvM
đPaper arxiv.org/pdf/2307.06737.pdf
đProject mirapurkrabek.github.io/RePoGen-paper
đCode github.com/MiraPurkrabek/RePoGen
đRePoGen: novel synthetic data generator of extreme/realistic poses of humans
đReview https://t.ly/ecBvM
đPaper arxiv.org/pdf/2307.06737.pdf
đProject mirapurkrabek.github.io/RePoGen-paper
đCode github.com/MiraPurkrabek/RePoGen
đĨ12đ2đ1đ¤¯1đą1
This media is not supported in your browser
VIEW IN TELEGRAM
đĄ DATID-3D: Text-to-3D Generation đĄ
đ A novel domain adaptation method for 3D via text-to-image diffusion. đ¤-Demo available!
đReview https://t.ly/TCL-B
đPaper arxiv.org/pdf/2211.16374.pdf
đProject gwang-kim.github.io/datid_3d/
đCode github.com/gwang-kim/DATID-3D
đ¤ huggingface.co/spaces/gwang-kim/DATID-3D
đColab colab.research.google.com/drive/1e9NSVB7x_hjz-nr4K0jO4rfTXILnNGtA?usp=sharing
đ A novel domain adaptation method for 3D via text-to-image diffusion. đ¤-Demo available!
đReview https://t.ly/TCL-B
đPaper arxiv.org/pdf/2211.16374.pdf
đProject gwang-kim.github.io/datid_3d/
đCode github.com/gwang-kim/DATID-3D
đ¤ huggingface.co/spaces/gwang-kim/DATID-3D
đColab colab.research.google.com/drive/1e9NSVB7x_hjz-nr4K0jO4rfTXILnNGtA?usp=sharing
đ¤¯5
This media is not supported in your browser
VIEW IN TELEGRAM
đ§¯Neural Focal Modulation VARđ§¯
đA novel architecture for video recognition that models both local/global context
đReview https://t.ly/rF_fk
đPaper arxiv.org/pdf/2307.06947.pdf
đProject talalwasim.github.io/Video-FocalNets
đCode github.com/TalalWasim/Video-FocalNets
đA novel architecture for video recognition that models both local/global context
đReview https://t.ly/rF_fk
đPaper arxiv.org/pdf/2307.06947.pdf
đProject talalwasim.github.io/Video-FocalNets
đCode github.com/TalalWasim/Video-FocalNets
đĨ8âĄ1đ1đ¤Š1
This media is not supported in your browser
VIEW IN TELEGRAM
đ Gen-AI as representation learner đ
đDreamTeacher: novel self-supervised feats. representation learning framework that utilizes gen-nets for pre-training downstream image backbones
đReview https://t.ly/RL8iG
đPaper arxiv.org/pdf/2307.07487.pdf
đProject research.nvidia.com/labs/toronto-ai/DreamTeacher
đDreamTeacher: novel self-supervised feats. representation learning framework that utilizes gen-nets for pre-training downstream image backbones
đReview https://t.ly/RL8iG
đPaper arxiv.org/pdf/2307.07487.pdf
đProject research.nvidia.com/labs/toronto-ai/DreamTeacher
đĨ9đ2đ¤¯1
This media is not supported in your browser
VIEW IN TELEGRAM
â #SelfDriving? It's all about weather! â
đNovel self-supervised MDE method to handle adverse weather in real-world autonomous driving
đReview https://t.ly/tcLQW
đPaper arxiv.org/pdf/2307.08357.pdf
đProject kieran514.github.io/Robust-Depth-Project/
đNovel self-supervised MDE method to handle adverse weather in real-world autonomous driving
đReview https://t.ly/tcLQW
đPaper arxiv.org/pdf/2307.08357.pdf
đProject kieran514.github.io/Robust-Depth-Project/
â¤7đ3đ¤¯1đą1