🍧Monocular 3D Clothed Human🍧

👉MultiGO++ is a novel framework for monocular 3D clothed human reconstruction via geometry-texture collaboration. New SOTA but no code announced🥲

👉Review https://t.ly/YKY44
👉Paper arxiv.org/pdf/2603.04993
👉Project 3dagentworld.github.io/multigo++

🎪SOTA Arbitrary Tracking🎪

👉TAPFormer is the novel SOTA transformer-based framework that performs asynchronous temporal-consistent fusion of frames and events for robust and high-freq point tracking. Repo & Dataset under MIT💙

👉Review https://t.ly/-q4wm
👉Paper https://arxiv.org/pdf/2603.04989
👉Project http://tapformer.github.io/
👉Repo https://github.com/ljx1002/TAPFormer

📊Real-Time Scene Graph📊

👉REACT++ by Umea University is the new state-of-the-art model for real-time SGG: 20% faster with a gain of 10% in relation prediction accuracy on average. Code under MIT💙

👉Review https://t.ly/c12VX
👉Paper https://arxiv.org/pdf/2603.06386
👉Repo https://github.com/Maelic/SGG-Benchmark

🔥Holistic 3D Spatial Intelligence🔥

👉Holi-Spatial is the first fully automated pipeline capable of converting raw video streams into holistic 3D spatial annotations without human intervention. Code/Data announced💙

👉Review https://t.ly/PDpr9
👉Paper https://lnkd.in/dTbMuZCm
👉Project https://lnkd.in/d66CYB4q
👉Repo https://lnkd.in/dAGzShXj

🍓Surface Light Tokenizer🍓

👉Apple unveils LITO a novel latent flow matching model enables HQ image-to-3D. Latent representation that encodes a surface light field into a compact set of latent vectors. Impressive results but no code🥲

👉Review https://t.ly/xcWNe
👉Paper https://lnkd.in/dYHwY4YX
👉Project https://lnkd.in/dtJT8bXy

☄️ OmniStream Backbone ☄️

👉Novel unified streaming visual backbone that effectively perceives, reconstructs, and acts from diverse visual inputs. Repo/Models announced💙

👉Review https://t.ly/_zZMO
👉Paper arxiv.org/pdf/2603.12265
👉Project go2heart.github.io/omnistream/
👉Repo github.com/Go2Heart/OmniStream

🌈 New SOTA Video Depth 🌈

👉DVD is the new Video Depth Estimation SOTA with full training suite available under Apache2.0💙

👉Review https://t.ly/gpCkG
👉Paper https://arxiv.org/pdf/2603.12250
👉Project https://dvd-project.github.io/
👉Repo github.com/EnVision-Research/DVD

🤖Physically-Plausible Human🤖

👉PhysMoDPO is a novel direct preference optimization framework for humanoid motion generation. Repo under MIT💙

👉Review https://t.ly/clf8w
👉Paper https://arxiv.org/pdf/2603.13228
👉Project https://mael-zys.github.io/PhysMoDPO/
👉Repo https://github.com/Mael-zys/PhysMoDPO

🍧10,000× faster SAM-3D🍧

👉Fast SAM 3D Body achieves up to 10.9× speedup, over 10,000× faster MHR-to-SMPL conversion -> real-time humanoid control from RGB. Repo available💙

👉Review https://t.ly/uHx84
👉Paper https://arxiv.org/pdf/2603.15603
👉Project yangtiming.github.io/Fast-SAM-3D-Body-Page/
👉Repo https://github.com/yangtiming/Fast-SAM-3D-Body

🍓Material-Aware Grouping🍓

👉Material Magic Wand (Adobe) is a tool for material-aware grouping of parts in untextured 3D meshes. Given one selected part, it automatically retrieves the other parts in the same shape by its material. Repo announced💙

👉Review https://t.ly/q00SU
👉Paper https://arxiv.org/pdf/2603.17370
👉Project umangi-jain.github.io/material-magic-wand/
👉Repo TBA

🦪OccAny: Universal 3D Occupancy🦪

👉OccAny by Valeo is a novel unified framework for generalized unconstrained urban 3D occupancy prediction. Repo under Apache 2.0💙

👉Review https://t.ly/FFiU0
👉Paper https://arxiv.org/pdf/2603.23502
👉Project https://valeoai.github.io/OccAny/
👉Repo https://github.com/valeoai/OccAny

🐍Pose-Appearance-Motion for HOI🐍

👉PAM is a novel Pose–Appearance–Motion Engine for controllable Hand–Object Interaction SOTA video generation. Repo/models available💙

👉Review https://t.ly/JU4MD
👉Paper arxiv.org/pdf/2603.22193
👉Project gasaiyu.github.io/PAM.github.io/
👉Repo https://github.com/GasaiYU/PAM

💥 GaussianGPT 3D GSC💥

👉From TUM, GaussianGPT: transformer-based 3D Gaussians generation via next-token prediction -> full 3D complex indoor scene. Repo announced💙

👉Review https://t.ly/bj-lL
👉Paper arxiv.org/pdf/2603.26661
👉Project nicolasvonluetzow.github.io/GaussianGPT/
👉Repo TBA

👌HandX: Scaling Hands Motion👌

👉 HandX is a unified foundation spanning data, annotation, and evaluation: novel large-scale dataset of bimanual & dexterous motions with fine-grained textual. Around 6M frames. Repo available💙

👉Review https://t.ly/1nGxw
👉Paper https://arxiv.org/pdf/2603.28766
👉Project https://handx-project.github.io/
👉Repo github.com/handx-project/HandX

🌵SOTA Training-Free In-Context Segmentation🌵

👉INSID3 is the new SOTA, training-free approach that segments concepts at varying granularities only from frozen DINOv3 features, given an in-context example. Repo under Apache 2.0💙

👉Review https://t.ly/NVWHN
👉Paper arxiv.org/pdf/2603.28480
👉Project visinf.github.io/INSID3/
👉Repo github.com/visinf/INSID3

🪬Camera Raw Image Generation🪬

👉RawGen by #Samsung is a generative approach that learns the complex distribution of raw sensor data directly, enabling high-fidelity generation from either text descriptions or standard sRGB images across arbitrary camera sensors. Linear raw image once, then apply any ISP operation. Repo announced💙

👉Review https://t.ly/_QVKP
👉Paper https://arxiv.org/pdf/2604.00093
👉Project https://dy112.github.io/rawgen-page/
👉Repo TBA

🍎Video Object Deletion🍎

👉Void by Netflix is a novel video object removal framework designed to perform physically-plausible inpainting in very complex scenarios. Repo under Apache 2.0💙

👉Review https://t.ly/cMVny
👉Paper https://arxiv.org/pdf/2604.02296
👉Project https://void-model.github.io/
👉Repo https://github.com/Netflix/void-model

🔥Vanast: VTON w/ Human Animation🔥

👉SNU unveils a novel unified framework that generates garment-transferred human animation videos directly from a single human/garment images, and pose guidance clip. Repo announced💙

👉Review https://t.ly/c0t79
👉Paper arxiv.org/pdf/2604.04934
👉Project hyunsoocha.github.io/vanast/
👉Repo github.com/snuvclab/vanast