Loading video...

Video Failed to Load

Go Home

๐Ÿ“ข๐Ÿ“ข๐Ÿ“ข NoKSR: Kernel-Free Neural Surface Reconstruction via Point Cloud Serialization #3DV2025 TL;DR: neural reconstruction with a simpler architecture (no linear systems to solve), and up to 3x speedup vs. voxel-based methods!

taiyasaki's profile picture

Andrea Tagliasacchi @CVPR

16,852 subscribers

16,962 views โ€ข 1 year ago โ€ขvia X (Twitter)

Science & TechnologyEducation#3DV2025
LIVE

Anya Rossiโ€ข Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

๐Ÿ“ขDNF: Generating 4D animations with dictionary-based neural fields! Xinyi Zhang presents a new dictionary-based neural field for unconditional 4D generation of deforming shapes -- generating motions with high-quality shape and temporal consistency.
0:19

๐Ÿ“ขDNF: Generating 4D animations with dictionary-based neural fields! Xinyi Zhang presents a new dictionary-based neural field for unconditional 4D generation of deforming shapes -- generating motions with high-quality shape and temporal consistency.

Angela Dai

14,639 views โ€ข 1 year ago

(1/2) ๐Ÿ“ข๐๐๐†๐€: ๐๐ž๐ฎ๐ซ๐š๐ฅ ๐๐š๐ซ๐š๐ฆ๐ž๐ญ๐ซ๐ข๐œ ๐†๐š๐ฎ๐ฌ๐ฌ๐ข๐š๐ง ๐€๐ฏ๐š๐ญ๐š๐ซ๐ฌ ๐Ÿ“ข #SIGGRAPHAsia We leverage a neural parametric representation to facilitate precise control over 3D Gaussians to obtain high-fidelity avatars.
0:19

(1/2) ๐Ÿ“ข๐๐๐†๐€: ๐๐ž๐ฎ๐ซ๐š๐ฅ ๐๐š๐ซ๐š๐ฆ๐ž๐ญ๐ซ๐ข๐œ ๐†๐š๐ฎ๐ฌ๐ฌ๐ข๐š๐ง ๐€๐ฏ๐š๐ญ๐š๐ซ๐ฌ ๐Ÿ“ข #SIGGRAPHAsia We leverage a neural parametric representation to facilitate precise control over 3D Gaussians to obtain high-fidelity avatars.

Matthias Niessner

30,741 views โ€ข 1 year ago

๐Ÿ“ขPix2NPHM: Learning to Regress NPHM Reconstructions From a Single Image๐Ÿ“ข We directly regress neural parametric head models (NPHMs) from a single image โ€” fast, stable, and significantly more expressive than classical 3DMMs such as FLAME. Face tracking & 3D reconstruction are often limited by the representational capacity of PCA-based face models. By lifting NPHMs to a first-class reconstruction primitive, we enable more accurate geometry, richer expressions, and finer animation control. Pix2NPHM obtains fast and reliable NPHM reconstructions on real-world data. Inference-time optimization against surface normals and canonical point maps can further increase fidelity. Key to successful and generalized training of our ViT-based network are: (1) large-scale registration of existing 3D head datasets, and (2) self-supervised training on vast in-the-wild 2D video datasets using pseudo ground-truth surface normals. Finally, we show that geometry-aware pretraining on pixel-aligned reconstruction tasks significantly outperforms generic visual pretraining (e.g., DINO-style features) in terms of generalization. ๐ŸŒ ๐ŸŽฅ Great work by Simon Giebenhain, Tobias Kirschstein, Liam Schoneveld, Davide Davoli, Zhe Chen
0:17

๐Ÿ“ขPix2NPHM: Learning to Regress NPHM Reconstructions From a Single Image๐Ÿ“ข We directly regress neural parametric head models (NPHMs) from a single image โ€” fast, stable, and significantly more expressive than classical 3DMMs such as FLAME. Face tracking & 3D reconstruction are often limited by the representational capacity of PCA-based face models. By lifting NPHMs to a first-class reconstruction primitive, we enable more accurate geometry, richer expressions, and finer animation control. Pix2NPHM obtains fast and reliable NPHM reconstructions on real-world data. Inference-time optimization against surface normals and canonical point maps can further increase fidelity. Key to successful and generalized training of our ViT-based network are: (1) large-scale registration of existing 3D head datasets, and (2) self-supervised training on vast in-the-wild 2D video datasets using pseudo ground-truth surface normals. Finally, we show that geometry-aware pretraining on pixel-aligned reconstruction tasks significantly outperforms generic visual pretraining (e.g., DINO-style features) in terms of generalization. ๐ŸŒ ๐ŸŽฅ Great work by Simon Giebenhain, Tobias Kirschstein, Liam Schoneveld, Davide Davoli, Zhe Chen

Matthias Niessner

37,827 views โ€ข 6 months ago

Introducing Neural Developmental Programs (NDPs)๐Ÿงฌ๐Ÿง Instead of neural networks with fixed architectures, we allow neural networks to grow through a dynamic self-organizing process, inspired by how biological nervous systems develop๐Ÿ‘‡ PDF:
0:19

Introducing Neural Developmental Programs (NDPs)๐Ÿงฌ๐Ÿง Instead of neural networks with fixed architectures, we allow neural networks to grow through a dynamic self-organizing process, inspired by how biological nervous systems develop๐Ÿ‘‡ PDF:

Sebastian Risi

239,469 views โ€ข 2 years ago

SplatVoxel: History-Aware Novel View Streaming without Temporal Training Contributions: โ€ข We propose a hybrid Splat-Voxel feed-forward reconstruction framework that leverages historical information to enable novel view streaming, without relying on multi-view video datasets for training. โ€ข We develop an efficient sparse voxel transformer with a coarse-to-fine voxel representation, outperforming existing feed-forward Gaussian splatting methods. โ€ข Experiment results demonstrate that our proposed framework enhances novel view synthesis for streaming scene reconstruction, providing better visual quality and reduced temporal artifacts through history-aware modeling.
0:18

SplatVoxel: History-Aware Novel View Streaming without Temporal Training Contributions: โ€ข We propose a hybrid Splat-Voxel feed-forward reconstruction framework that leverages historical information to enable novel view streaming, without relying on multi-view video datasets for training. โ€ข We develop an efficient sparse voxel transformer with a coarse-to-fine voxel representation, outperforming existing feed-forward Gaussian splatting methods. โ€ข Experiment results demonstrate that our proposed framework enhances novel view synthesis for streaming scene reconstruction, providing better visual quality and reduced temporal artifacts through history-aware modeling.

MrNeRF

10,823 views โ€ข 1 year ago

NO FUCKING WAY SEUNGRI!!!!! #SEUNGRI NEW ๐Ÿ“ข ๐Ÿ“ข ๐Ÿ“ข ๐Ÿ“ข ๐Ÿ“ข
0:10

NO FUCKING WAY SEUNGRI!!!!! #SEUNGRI NEW ๐Ÿ“ข ๐Ÿ“ข ๐Ÿ“ข ๐Ÿ“ข ๐Ÿ“ข

๐Ÿฆ‹๐ŸผSeungri is my Princeยณโต๐Ÿ’™๐Ÿผ

17,256 views โ€ข 4 months ago

NN-SVG: Create neural network architecture drawings parametrically! Export them to SVG files and use them in your work!
0:17

NN-SVG: Create neural network architecture drawings parametrically! Export them to SVG files and use them in your work!

Daily Dose of Data Science

109,708 views โ€ข 9 months ago

NN-SVG: Create neural network architecture drawings parametrically! Export them to SVG files and use them in your work!
0:17

NN-SVG: Create neural network architecture drawings parametrically! Export them to SVG files and use them in your work!

Daily Dose of Data Science

70,805 views โ€ข 1 year ago

๐Ÿ“ข๐Ÿ“ขOpen dOTC is LIVE! ๐Ÿ“ข๐Ÿ“ข ๐Ÿ“ˆ Head to the platform to start trading and remember trading fees automatically buy and burn $SMT! ๐Ÿ‘‰ #tokenization #RWAs #permissionlessRWAs
0:10

๐Ÿ“ข๐Ÿ“ขOpen dOTC is LIVE! ๐Ÿ“ข๐Ÿ“ข ๐Ÿ“ˆ Head to the platform to start trading and remember trading fees automatically buy and burn $SMT! ๐Ÿ‘‰ #tokenization #RWAs #permissionlessRWAs

Swarm

17,214 views โ€ข 2 years ago

Nuvo: Neural UV Mapping! It's super difficult to UV map/texture atlas geometry produced by 3D reconstruction and generation pipelines. Nuvo works on all kinds of "unruly" 3D representations (NeRF, DreamFusion, etc.) and enables easy appearance editing! 1/3
0:12

Nuvo: Neural UV Mapping! It's super difficult to UV map/texture atlas geometry produced by 3D reconstruction and generation pipelines. Nuvo works on all kinds of "unruly" 3D representations (NeRF, DreamFusion, etc.) and enables easy appearance editing! 1/3

Pratul Srinivasan

20,842 views โ€ข 2 years ago

Human Hair Reconstruction with Strand-Aligned 3D Gaussians Contributions (cited): โ€“ We propose a new 3D line lifting scheme that uses a modified 3DGS reconstruction technique to lift 2D orientation maps into a 3D field while also providing refinement of the camera parameters; โ€“ We introduce a dual representation of hair strand polylines and 3D Gaussians to achieve differentiable rasterization of hair strands and leverage photometric constraints for strand-based hair reconstruction; โ€“ Based on these components, we propose a coarse-to-fine optimization method for prior-guided hair reconstruction that leverages both latent and explicit representations of the hairstyle.
0:13

Human Hair Reconstruction with Strand-Aligned 3D Gaussians Contributions (cited): โ€“ We propose a new 3D line lifting scheme that uses a modified 3DGS reconstruction technique to lift 2D orientation maps into a 3D field while also providing refinement of the camera parameters; โ€“ We introduce a dual representation of hair strand polylines and 3D Gaussians to achieve differentiable rasterization of hair strands and leverage photometric constraints for strand-based hair reconstruction; โ€“ Based on these components, we propose a coarse-to-fine optimization method for prior-guided hair reconstruction that leverages both latent and explicit representations of the hairstyle.

MrNeRF

106,525 views โ€ข 1 year ago

#AA23 with Prasanth Neel ๐Ÿ“ข #AA23 with Boyapati sreenu ๐Ÿ“ข #AA23 with Mehar Ramesh ๐Ÿ“ข #AA23 with Linguswami ๐Ÿ“ข #AA23 with Vetrimaran ๐Ÿ“ข #AA23 with Lokesh kanakaraju ๐Ÿ“ข #AA23 with SanjayLeelaBansali ๐Ÿ“ข #AA23 with joseph ๐Ÿ“ข Okka movie ni intha manditho chesthunnavante Allu Arjun ๐Ÿ™๐Ÿป
0:11

#AA23 with Prasanth Neel ๐Ÿ“ข #AA23 with Boyapati sreenu ๐Ÿ“ข #AA23 with Mehar Ramesh ๐Ÿ“ข #AA23 with Linguswami ๐Ÿ“ข #AA23 with Vetrimaran ๐Ÿ“ข #AA23 with Lokesh kanakaraju ๐Ÿ“ข #AA23 with SanjayLeelaBansali ๐Ÿ“ข #AA23 with joseph ๐Ÿ“ข Okka movie ni intha manditho chesthunnavante Allu Arjun ๐Ÿ™๐Ÿป

เฐ…เฐฎเฐพเฐฏเฐ•เฑเฐกเฑ โ„ข

93,573 views โ€ข 1 year ago

DLSS 4.5 is coming to Ray Reconstruction! Massive image quality improvements for games with Ray Reconstruction ๐Ÿ‘
0:11

DLSS 4.5 is coming to Ray Reconstruction! Massive image quality improvements for games with Ray Reconstruction ๐Ÿ‘

Jacob Freeman

83,894 views โ€ข 1 month ago

๐Ÿš€ Excited to share our #ICLR2025 work on planning with neural dynamics models! While our lab has developed diverse neural dynamics models for manipulating rigid, deformable, and granular objects, having the model alone doesnโ€™t solve the problemโ€”planning with it remains a challenge. ๐Ÿ’ก Enter BaB-ND, led by Keyi and Jiangwei! We propose a scalable, GPU-accelerated branch-and-bound algorithm, inspired by neural network verification, to enable effective planning for diverse objects modeled with neural dynamics. ๐Ÿ”— Project page (open-source + detailed docs!): ๐ŸŽฅ Watch the video to see T being pushed around obstacles, and check out Keyiโ€™s thread for more details!
0:14

๐Ÿš€ Excited to share our #ICLR2025 work on planning with neural dynamics models! While our lab has developed diverse neural dynamics models for manipulating rigid, deformable, and granular objects, having the model alone doesnโ€™t solve the problemโ€”planning with it remains a challenge. ๐Ÿ’ก Enter BaB-ND, led by Keyi and Jiangwei! We propose a scalable, GPU-accelerated branch-and-bound algorithm, inspired by neural network verification, to enable effective planning for diverse objects modeled with neural dynamics. ๐Ÿ”— Project page (open-source + detailed docs!): ๐ŸŽฅ Watch the video to see T being pushed around obstacles, and check out Keyiโ€™s thread for more details!

Yunzhu Li

10,561 views โ€ข 1 year ago

Chciaล‚abym, to jakoล› skomentowaฤ‡ ๐Ÿ“ขโ€ผ๏ธ๐Ÿ“ขโ€ผ๏ธ ale kolejny raz mnie zatkaล‚o๐Ÿ“ข๐Ÿ˜ก๐Ÿ“ข๐Ÿ˜ก Zostawiam Wam๐Ÿคท๐Ÿ“ข๐Ÿคท
0:19

Chciaล‚abym, to jakoล› skomentowaฤ‡ ๐Ÿ“ขโ€ผ๏ธ๐Ÿ“ขโ€ผ๏ธ ale kolejny raz mnie zatkaล‚o๐Ÿ“ข๐Ÿ˜ก๐Ÿ“ข๐Ÿ˜ก Zostawiam Wam๐Ÿคท๐Ÿ“ข๐Ÿคท

Monia

23,920 views โ€ข 4 months ago

๐Ÿ“ข Impossible Cloud Network's Node Sale is coming in December! ๐Ÿ“ข The countdown begins! This is your chance to become a node operator with ICN, helping to build a secure, decentralized cloud infrastructure while earning rewards for your contribution. More info on the Node Sale will be released soon, so stay tuned! ๐Ÿš€
0:12

๐Ÿ“ข Impossible Cloud Network's Node Sale is coming in December! ๐Ÿ“ข The countdown begins! This is your chance to become a node operator with ICN, helping to build a secure, decentralized cloud infrastructure while earning rewards for your contribution. More info on the Node Sale will be released soon, so stay tuned! ๐Ÿš€

Impossible Cloud Network

91,333 views โ€ข 1 year ago

AGOTT ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ
0:15

AGOTT ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ๐Ÿ—ฃ๏ธ๐Ÿ“ข๐Ÿ’ฅ

๐Ÿฅฅ

30,998 views โ€ข 4 months ago