AI with Papers - Artificial Intelligence & Deep Learning
17.3K subscribers
165 photos
291 videos
12 files
1.47K links
All the AI with papers. Every day fresh updates about #DeepLearning #MachineLearning #LLM & #ComputerVision

Curated by Alessandro Ferrari | https://www.linkedin.com/in/visionarynet/

#AI #chatGPT
Download Telegram
This media is not supported in your browser
VIEW IN TELEGRAM
🐇Segment where you look🐇

👉KAUST (+Adobe) revisited the interface through which a vision-language model passes the location of its referent to a segmentation decoder. Code announced💙

👉Review https://lnkd.in/p/er_ifxGR
👉Paper https://lnkd.in/eATXw98R
👉Project https://lnkd.in/eXSWvxPN
👉Repo TBA
❤4👍3🔥3
This media is not supported in your browser
VIEW IN TELEGRAM
🧊SOTA category-level 9D Pose🧊

👉SJTU at #NeurIPS 2026 with the new SOTA among RGB-based methods for category-level 9D pose estimation. Code announced💙

👉Review https://lnkd.in/p/e5BvM7Ua
👉Paper https://arxiv.org/pdf/2610.18374
👉Project https://souglad21.github.io/SchemaPose/
👉Repo TBA
❤7👍2👏1
This media is not supported in your browser
VIEW IN TELEGRAM
🎊 Pixel-Level 4D Humans 🎊

👉Dense4D from the Imperial College is the first feed-forward model dedicated to dense 4D correspondence across the body, clothing and hair from a set of monocular RGB images, without a body template. Code announced💙

👉Review https://lnkd.in/p/e9GXGggz
👉Paper https://arxiv.org/pdf/2610.22325
👉Project https://alpguler.com/dense4d/
👉Repo TBA
❤8🔥6👍3👏1
🔥700 FPS Humans Recovery🔥

👉One model, any sparse signal: OmniPoser recovers full-body 3D human motion from 3D keypoints, sparse IMU rotations, head-mounted 6-DoF poses, or any subset or combination of these signals. All within a SINGLE model. Repo under nc-cc license💙

👉Review https://lnkd.in/p/et-Pk-Dy
👉Paper https://arxiv.org/pdf/2610.19857
👉Project https://jamesbridgewater51.github.io/omniposer-eccv/
👉Repo https://github.com/JamesBridgewater51/omniposer
1❤5🔥4👏1
This media is not supported in your browser
VIEW IN TELEGRAM
🏵️Lang-Informed Lossless Encoder🏵️

👉Stanford's LILIE: an invertible image transform trained such that a vision-language model can recover captions from noised features. Code, encoders, and generative models💙

👉Review https://lnkd.in/p/dq3SK_nc
👉Paper https://arxiv.org/pdf/2610.23653
👉Project https://www.suchiragarwal.com/lilie/
👉Repo https://github.com/SuchirAgarwal/lilie
❤4👏2👍1
This media is not supported in your browser
VIEW IN TELEGRAM
🔥SOTA Video Matting🔥

👉VideoMaMa++ (#NeurIPS 2026): mask-guided video matting framework that recovers fine-grained details (e.g., hair strands) and scales reliably to long clips. Impressive, code announced💙

👉Review https://lnkd.in/p/dj3_zE3P
👉Paper https://arxiv.org/pdf/2610.26857
👉Project https://cvlab-kaist.github.io/VideoMaMa-PlusPlus/
👉Repo https://github.com/cvlab-kaist/VideoMaMa-PlusPlus
❤6👍2🔥2👏1🤯1
This media is not supported in your browser
VIEW IN TELEGRAM
🥭Agentic Manga->Animation🥭

👉Manga2Animate: turning existing manga pages into coherent animation that stays faithful to the original manga. Impressive results, best with sound. Code released.💙

👉Review https://lnkd.in/p/dHpdPAj3
👉Paper https://arxiv.org/pdf/2610.26851
👉Project https://annchenn.github.io/Manga2Animate/
👉Repo https://github.com/annchenn/Manga2Animate
🔥20🤯8❤5👍3😍2💩1
This media is not supported in your browser
VIEW IN TELEGRAM
🍵3D from Single Human RGB🍵

👉Real-to-semantic-to-real pipeline that turns one monocular RGB video of a human demonstration into thousands of robot demonstrations. Code TBA💙

👉Review https://lnkd.in/p/dt-B2QMt
👉Paper https://arxiv.org/pdf/2610.28880
👉Project https://real2sem2real.github.io/
👉Repo TBA
🔥7❤1👍1
This media is not supported in your browser
VIEW IN TELEGRAM
🧊 SOTA Foundation Pose 🧊

👉From one segmented RGB-D observation, directly predict 6D pose, metric size, and camera-frame completed shape. Repo under Apace 2.0💙

👉Review https://lnkd.in/p/dXpXt7Pz
👉Paper https://arxiv.org/pdf/2610.31623
👉Project https://jrryzh.github.io/F2G-Pose/
👉Repo https://github.com/jrryzh/F2G-Pose
❤7🤯2🔥1👏1
This media is not supported in your browser
VIEW IN TELEGRAM
🫀 SKIM V2 is out🫀

👉SOMA recovers muscle deformation over time from surface observations captured by multi-view RGB cameras, the first approach of this kind. SKIM v2: a new soft-tissue deformation dataset. Code under MIT, data & weights NC💙

👉Review https://lnkd.in/p/dB4XS7vB
👉Paper https://lnkd.in/d64S_qxj
👉Project https://lnkd.in/dWtavYGn
👉Repo https://lnkd.in/dXNJqY6N
🔥3❤2👏1😍1
This media is not supported in your browser
VIEW IN TELEGRAM
🍏 70+ FPS 1080p Matting 🍏

👉MatAnyone-E: on a single H200: 73.07 FPS at 1080p and 28.62 FPS at 4K. Code TBA, S-LAB license💙

👉Review https://lnkd.in/p/deB-YcP9
👉Paper https://arxiv.org/pdf/2610.35724
👉Project https://shuyun-wang.github.io/projects/MatAnyone-E/
👉Repo https://github.com/Shuyun-Wang/MatAnyone-E
❤3🔥2👏1
This media is not supported in your browser
VIEW IN TELEGRAM
🦖MoCap Anything V2 is out🦖

👉End-to-end motion capture for arbitrary skeletons: from a monocular video plus a reference asset straight to BVH-ready joint rotations, with no analytical IK. Works on humans, animals and custom rigs. Repo & weights under MIT, but RMBG-1.4 background remover is non-commercial💙

👉Review https://lnkd.in/p/dtZjGwh7
👉Paper https://lnkd.in/dG-dtCFT
👉Project https://lnkd.in/d7aGbYPj
👉Repo https://lnkd.in/dRWZNKjY
❤8🔥5👍1👏1
This media is not supported in your browser
VIEW IN TELEGRAM
🥭Fast/Accurate Hand Estimation🥭

👉FAH: 2D/3D joints and surface points, fit MANO or MHR-Hand, and optionally estimate anatomical parts and contact without object geometry at inference. Repo under Apache 2.0💙

👉Review https://lnkd.in/p/dfTBMFA8
👉Paper https://arxiv.org/pdf/2611.07711
👉Project https://www.jinkuncao.com/FAH/
👉Repo https://github.com/noahcao/FAH
❤3👍1🔥1👏1
This media is not supported in your browser
VIEW IN TELEGRAM
👑VINO: 3D w/ Volume Transformers👑

👉VINO revisits the previous key design in 3D Representation: building on Volume Transformer, VINO constructs reliable targets using clean global views and a target head decoupled from the student. Repo announced💙

👉Review https://lnkd.in/p/dyWDkewj
👉Paper https://arxiv.org/pdf/2611.07176
👉Project https://adriankruse.github.io/VINO/
👉Repo https://github.com/AdrianKruse/VINO
🔥5❤4👏1
This media is not supported in your browser
VIEW IN TELEGRAM
🪂 SOTA Dense 3D Tracking 🪂

👉Technion & Nvidia unveil ReLift, a novel/efficient feed-forward net that recasts dense 3D tracking as geometric trajectory repair and completion. Repo under MIT💙

👉Review https://lnkd.in/p/d82m2pDJ
👉Paper https://arxiv.org/pdf/2610.42136
👉Project https://relift3d.github.io/
👉Repo https://github.com/ReLift3D/ReLift
🔥5❤2👏1
This media is not supported in your browser
VIEW IN TELEGRAM
🍵Map2SLAM: real time SLAM🍵

👉Map2SLAM is the new SOTA on KITTI, VBR, and Oxford Spires, reducing mean ATE by over 40% compared with the best uncalibrated baselines, while maintaining real-time tracking at 28+ FPS at 518×154 on an NVIDIA RTX 5090D GPU. Repo announced💙

👉Review https://lnkd.in/p/dM8zEYPH
👉Paper https://arxiv.org/pdf/2611.11169
👉Project https://robbyant-research.github.io/Map2SLAM/
👉Repo https://github.com/robbyant-research/Map2SLAM
❤11🔥3👏1🤣1
This media is not supported in your browser
VIEW IN TELEGRAM
🏜️Unified Generative Geometry🏜️

👉A single model jointly generates RGB video, depth, and camera geometry, turning visual generation and geometric prediction into conditional completion of the same 4D scene. Repo under MIT💙

👉Review https://lnkd.in/p/dskPQkQY
👉Paper https://lnkd.in/dxHWWHsX
👉Project https://lnkd.in/dvUd5ZzN
👉Repo https://lnkd.in/dRJuQzqU
❤2🔥1🤯1
This media is not supported in your browser
VIEW IN TELEGRAM
🍏 MoviEdit: easy editing 🍏

👉MoviEdit by Adobe is an end-to-end framework for scalable long-form multi-shot video editing that jointly addresses the challenges of data, modeling, and evaluation. Impressive results💙

👉Review https://lnkd.in/p/d233HiA7
👉Paper https://arxiv.org/pdf/2611.02737
👉Project https://multi-shot-edit.github.io/
👉Repo https://github.com/adobe-research/MoviEdit-Bench
❤6🔥2👏1
This media is not supported in your browser
VIEW IN TELEGRAM
🟰Connected-Brick Reconstruction🟰

👉BrickFlow is an SE(3) flow-matching model trained to take in a known set of LEGO pieces (pile) with noisy poses (translucent) and denoise them into some valid structure. Code in a few days💙

👉Review https://lnkd.in/p/dc-wq3k5
👉Paper https://arxiv.org/pdf/2611.17321
👉Project https://kulits.github.io/BrickFlow/
👉Repo https://github.com/kulits/BrickFlow
🤯7❤3🔥1👏1
This media is not supported in your browser
VIEW IN TELEGRAM
🛏️ Spatial Foundry Dataset 🛏️

👉Spatial Foundry by Ranolabs contains 5,280 furnished residential 3D scenes with point clouds, editable meshes, aligned perspective and panoramic RGB, metric depth, 3D bounding boxes, and camera trajectories paired with video💙

👉Review https://lnkd.in/p/dkbFkjJp
👉Paper https://arxiv.org/pdf/2611.17318
👉Project https://ranolabs.com/Blog/spatial-foundry
❤3👍1🔥1