AI with Papers - Artificial Intelligence & Deep Learning
15K subscribers
95 photos
235 videos
11 files
1.26K links
All the AI with papers. Every day fresh updates on Deep Learning, Machine Learning, and Computer Vision (with Papers).

Curated by Alessandro Ferrari | https://www.linkedin.com/in/visionarynet/
Download Telegram
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿ—ฝ Neuralangelo Digital Twins. INSANE๐Ÿ—ฝ

๐Ÿ‘‰ A novel framework from #Nvidia for Hi-Fi 3D Digital twins.

๐Ÿ˜ŽReview https://t.ly/rxoF4
๐Ÿ˜ŽProject research.nvidia.com/labs/dir/neuralangelo
๐Ÿ˜ŽPaper research.nvidia.com/labs/dir/neuralangelo/paper.pdf
๐Ÿ”ฅ15๐Ÿ‘4๐Ÿคฏ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿฆœ ColorDiffuser: Text-to-Video Colorization ๐Ÿฆœ

๐Ÿ‘‰HK University unveils ColorDiffuser: adapting pre-trained text-to-image latent diffusion model for video colorization

๐Ÿ˜ŽReview https://t.ly/XGv_
๐Ÿ˜ŽPaper arxiv.org/pdf/2306.01732.pdf
๐Ÿ˜ŽProject colordiffuser.github.io/
๐Ÿ˜ŽCode github.com/ColorDiffuser/ColorDiffuser
๐Ÿคฏ8โค2๐Ÿคฉ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐ŸŒป Extending Mona Lisa with AI ๐ŸŒป

๐Ÿ‘‰ A guy on Reddit extends Mona Lisa Painting with #Photoshop AI. The result is surprising.

๐Ÿ˜ŽMore https://t.ly/j_2r
๐Ÿคฏ20๐Ÿ‘5๐Ÿคฉ4๐Ÿ”ฅ3๐Ÿ˜ฑ2๐Ÿคฃ2โšก1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿธ Segment Anything in HQ ๐Ÿธ

๐Ÿ‘‰HQ-SAM: SAM with the ability to accurately segment objects, maintaining promptable design, efficiency, zero-shot generalizability

๐Ÿ˜ŽReview https://t.ly/GxX5B
๐Ÿ˜ŽPaper arxiv.org/pdf/2306.01567.pdf
๐Ÿ˜ŽModels github.com/SysCV/SAM-HQ
๐Ÿ”ฅ18๐Ÿ‘4๐Ÿคฏ1๐Ÿ˜ฑ1๐Ÿ˜1
This media is not supported in your browser
VIEW IN TELEGRAM
๐ŸŒˆ Track Everything Everywhere ๐ŸŒˆ

๐Ÿ‘‰#Google unveils OmniMotion: full-length motion tracking for every pixel in every frame of video.

๐Ÿ˜ŽReview https://t.ly/Krvw
๐Ÿ˜ŽPaper arxiv.org/pdf/2306.05422.pdf
๐Ÿ˜ŽProject omnimotion.github.io/
๐Ÿ˜ŽDemo omnimotion.github.io/#interactive_demo
๐Ÿ˜ŽCode github.com/qianqianwang68/omnimotion
๐Ÿ”ฅ23โค5๐Ÿคฏ3๐Ÿคฉ1๐Ÿ’ฉ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿ‘๏ธ Scene Five: Through Her Eyes ๐Ÿ‘๏ธ

๐Ÿ‘‰ #3D scene reconstruction of what a person is observing using only the reflections of their eyes

๐Ÿ˜ŽReview https://t.ly/uBO6
๐Ÿ˜ŽPaper arxiv.org/pdf/2306.09348.pdf
๐Ÿ˜ŽProject https://world-from-eyes.github.io/
๐Ÿคฏ28๐Ÿ”ฅ12๐Ÿ’ฉ2๐Ÿคฉ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿงฟ NeRF-Supervised Deep Stereo ๐Ÿงฟ

๐Ÿ‘‰A novel pioneering pipeline for training deep stereo networks WITH NO ground-truth

๐Ÿ˜ŽReview https://t.ly/c7j-
๐Ÿ˜ŽProject nerfstereo.github.io/
๐Ÿ˜ŽDataset https://amsacta.unibo.it/id/eprint/7218/
๐Ÿ˜ŽCode github.com/fabiotosi92/NeRF-Supervised-Deep-Stereo
๐Ÿ˜ŽPaper https://openaccess.thecvf.com/content/CVPR2023/papers/Tosi_NeRF-Supervised_Deep_Stereo_CVPR_2023_paper.pdf
๐Ÿฅฐ8๐Ÿคฉ3โค1๐Ÿ‘1๐Ÿ’ฉ1๐Ÿ˜1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿซฃ Text-Guided Adversarial Makeup ๐Ÿซฃ

๐Ÿ‘‰Novel facial privacy protection via adversarial latent codes. Makeup vs Face Recognition.

๐Ÿ˜ŽReview https://t.ly/pBCP
๐Ÿ˜ŽPaper arxiv.org/pdf/2306.10008.pdf
๐Ÿ˜ŽCode github.com/fahadshamshad/Clip2Protect
โค6๐Ÿ‘1๐Ÿ”ฅ1๐Ÿฅฐ1๐Ÿ’ฉ1
Media is too big
VIEW IN TELEGRAM
๐Ÿฆท Few-Shot Geometry-Aware Keypoints ๐Ÿฆท

๐Ÿ‘‰UBC (+Flawless AI) unveils the new SOTA in semantic keypoints localization. Suitable for faces, animals, cars, mouth, teeth & more

๐Ÿ˜ŽReview https://t.ly/-0qN
๐Ÿ˜ŽPaper arxiv.org/pdf/2303.17216.pdf
๐Ÿ˜ŽProject xingzhehe.github.io/FewShot3DKP/
๐Ÿคฏ10๐Ÿ‘4โค2โšก2๐Ÿ‘2๐Ÿคฉ2๐Ÿ”ฅ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿš” Fooling Neural Forensic Classifiers ๐Ÿš”

๐Ÿ‘‰Adversarial faces able to fool the forensic classifiers, while remaining undetectable by humans

๐Ÿ˜ŽReview https://t.ly/33Cc
๐Ÿ˜ŽPaper arxiv.org/pdf/2306.13091.pdf
๐Ÿ˜ŽProject koushiksrivats.github.io/face_attribute_attack
๐Ÿ˜ŽCode github.com/koushiksrivats/face_attribute_attack
๐Ÿ˜ข6โค4๐Ÿ‘2๐Ÿ˜ฑ2๐Ÿพ2๐Ÿ‘1๐Ÿคฏ1๐Ÿ˜1
panohead_overview-min.gif
24.3 MB
๐Ÿฅ PanoHead: 3D Full-Head Synthesis ๐Ÿฅ

๐Ÿ‘‰#ByteDance (+UW-M) unveils PanoHead: 360โ—ฆ view-consistent portraits from a single-view image

๐Ÿ˜ŽReview https://t.ly/MrLNR
๐Ÿ˜ŽPaper arxiv.org/pdf/2303.13071.pdf
๐Ÿ˜ŽProject sizhean.github.io/panohead
๐Ÿ˜ŽCode github.com/sizhean/panohead
๐Ÿ”ฅ7โค4๐Ÿคฏ3๐Ÿ˜ฑ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿ”ฎSAM-PT: Segment Anything+Tracking๐Ÿ”ฎ

๐Ÿ‘‰SAM-PT is the first method to utilize sparse point propagation for Video Object Segmentation (VOS).

๐Ÿ˜ŽReview https://t.ly/QLMG
๐Ÿ˜ŽPaper arxiv.org/pdf/2307.01197.pdf
๐Ÿ˜ŽProject www.vis.xyz/pub/sam-pt/
๐Ÿ˜ŽCode github.com/SysCV/sam-pt
๐Ÿ”ฅ14โค7๐Ÿคฏ3๐Ÿ‘1๐Ÿ˜ฑ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿชฉ DISCO: Human Dance Generation ๐Ÿชฉ

๐Ÿ‘‰NTU (+ #Microsoft) unveils DISCO: a big step towards the Human Dance Generation.

๐Ÿ˜ŽReview https://t.ly/cNGX
๐Ÿ˜ŽPaper arxiv.org/pdf/2307.00040.pdf
๐Ÿ˜ŽProject disco-dance.github.io/
๐Ÿ˜ŽCode github.com/Wangt-CN/DisCo
๐Ÿ”ฅ13๐Ÿฅฐ4๐Ÿ˜2โšก1๐Ÿ‘1๐Ÿพ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿ›ฃ๏ธ STAR.: 3D-tracking w/ attention paradigm ๐Ÿ›ฃ๏ธ

๐Ÿ‘‰#Mercedes STAR: e2e 3D object tracking that follows the tracking-by-attention paradigm

๐Ÿ˜ŽReview https://t.ly/JoGj
๐Ÿ˜ŽPaper arxiv.org/pdf/2306.17602.pdf
๐Ÿ˜ŽProject simondoll.github.io/publications/star_track
๐Ÿ‘14๐Ÿ”ฅ1๐Ÿฅฐ1๐Ÿ‘1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿก Text2Cinemagraphs: Cinemagraph from text ๐Ÿก

๐Ÿ‘‰CMU (+ #Snap) unveils a fully automated method for creating cinemagraphs from text descriptions

๐Ÿ˜ŽReview https://t.ly/BwZs6
๐Ÿ˜ŽPaper arxiv.org/pdf/2307.03190.pdf
๐Ÿ˜ŽProject text2cinemagraph.github.io/website
๐Ÿ˜ŽCode github.com/text2cinemagraph/text2cinemagraph
โค12๐Ÿคฏ3๐Ÿ˜ฑ1๐Ÿคฉ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿ”ฅTest-Time Training on fire ๐Ÿ”ฅ

๐Ÿ‘‰Extending the TTT to the streaming setting. Suitable for Panoptic, Instance & Colorization.

๐Ÿ˜ŽReview https://t.ly/eZYA
๐Ÿ˜ŽPaper arxiv.org/pdf/2307.05014.pdf
๐Ÿ˜ŽProject https://video-ttt.github.io/
๐Ÿ˜ŽCode github.com/renwang435/video-ttt-release
๐Ÿ”ฅ10๐Ÿ‘3โšก1๐Ÿคฏ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿƒ Deepfake via casual self-scan ๐Ÿƒ

๐Ÿ‘‰TAU presents a novel approach to reenact an ID using only a casual self-scan

๐Ÿ˜ŽReview https://t.ly/9T8Wi
๐Ÿ˜ŽPaper arxiv.org/pdf/2307.06307.pdf
๐Ÿ˜ŽProject arielazary.github.io/PGR
๐Ÿคฏ7๐Ÿ‘6โค5๐Ÿ”ฅ1๐Ÿ˜ฑ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐ŸŽช Extreme Human Pose Estimation ๐ŸŽช

๐Ÿ‘‰RePoGen: novel synthetic data generator of extreme/realistic poses of humans

๐Ÿ˜ŽReview https://t.ly/ecBvM
๐Ÿ˜ŽPaper arxiv.org/pdf/2307.06737.pdf
๐Ÿ˜ŽProject mirapurkrabek.github.io/RePoGen-paper
๐Ÿ˜ŽCode github.com/MiraPurkrabek/RePoGen
๐Ÿ”ฅ12๐Ÿ‘2๐Ÿ‘1๐Ÿคฏ1๐Ÿ˜ฑ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐Ÿ’ก DATID-3D: Text-to-3D Generation ๐Ÿ’ก

๐Ÿ‘‰ A novel domain adaptation method for 3D via text-to-image diffusion. ๐Ÿค—-Demo available!

๐Ÿ˜ŽReview https://t.ly/TCL-B
๐Ÿ˜ŽPaper arxiv.org/pdf/2211.16374.pdf
๐Ÿ˜ŽProject gwang-kim.github.io/datid_3d/
๐Ÿ˜ŽCode github.com/gwang-kim/DATID-3D
๐Ÿค— huggingface.co/spaces/gwang-kim/DATID-3D
๐Ÿ˜ŽColab colab.research.google.com/drive/1e9NSVB7x_hjz-nr4K0jO4rfTXILnNGtA?usp=sharing
๐Ÿคฏ5