Marwan Taher

71 posts

Marwan Taher

Marwan Taher

@marwan_ptr

PhD student at the Dyson Robotics Lab @ Imperial College London.

Katılım Nisan 2011
485 Takip Edilen432 Takipçiler
Sabitlenmiş Tweet
Marwan Taher
Marwan Taher@marwan_ptr·
How can we run reconstruction models like π³ and Depth Anything 3 in real-time? We present KV-Tracker, a training-free approach, for real-time tracking of scenes and objects. Achieving up to 30 FPS! With @alzugarayign, @makezur, @XinKong_IC and @AjdDavison
English
10
97
704
64.7K
Marwan Taher
Marwan Taher@marwan_ptr·
KV-Tracker enables object-level reconstruction and tracking when provided with an object mask. The KV-cache can be saved and used later without any special initialisation procedure.
English
2
0
19
2K
Marwan Taher
Marwan Taher@marwan_ptr·
How can we run reconstruction models like π³ and Depth Anything 3 in real-time? We present KV-Tracker, a training-free approach, for real-time tracking of scenes and objects. Achieving up to 30 FPS! With @alzugarayign, @makezur, @XinKong_IC and @AjdDavison
English
10
97
704
64.7K
Marwan Taher
Marwan Taher@marwan_ptr·
Per-frame geometry from π³ is split into primitives via segmentation and tracked over time using dense 2D point tracks. With a compact per primitive pose, geometry is densely aligned, stitching primitives to create a complete reconstruction of the observed scene components.
Kirill Mazur@makezur

Introducing 4D Primitive-Mâché (4DPM), a new method for replayable 4D reconstruction from monocular videos. We split dynamic scenes into 3D primitives and recover their motion. 4DPM can infer object positions even after they leave view. Joint work with @marwan_ptr @AjdDavison

English
0
1
9
536
Marwan Taher
Marwan Taher@marwan_ptr·
ACE-SLAM naturally handles loop closure without special treatment, robustly deals with dynamic objects, while remaining lightweight (small MLP) and computationally efficient—making this representation compelling for SLAM.
Ignacio Alzugaray@alzugarayign

Excited to present ACE-SLAM, the first neural SLAM to use Scene Coordinate Regression as an implicit map representation Efficient (real-time from live stream), compressive (neural maps <1MB) and robust to dynamic scenes With @marwan_ptr and @AjdDavison ialzugaray.github.io/ace-slam

English
0
2
11
964
Marwan Taher retweetledi
Xin Kong
Xin Kong@XinKong_IC·
🚀 Excited to share CausNVS: Autoregressive Multi-view Diffusion for Flexible 3D Novel View Synthesis! Let’s recon 3D world generatively. CausNVS handles any number of input views, synthesizes novel views autoregressively, enables interactive streaming and flexible N-to-M NVS.
English
2
18
106
10.2K
Marwan Taher retweetledi
Riku Murai
Riku Murai@rmurai0610·
Introducing MASt3R-SLAM, the first real-time monocular dense SLAM with MASt3R as a foundation. Easy to use like DUSt3R/MASt3R, from an uncalibrated RGB video it recovers accurate, globally consistent poses & a dense map. With @eric_dexheimer*, @AjdDavison (*Equal Contribution)
English
42
255
1.4K
203.2K
Ayoung
Ayoung@_ayoungk·
3D Gaussian Splatting SLAM demo!
English
9
35
392
63.5K
Samuel
Samuel@_Sir_MuEl·
@_ayoungk What's the name of this paper please?
English
1
0
1
563
Marwan Taher
Marwan Taher@marwan_ptr·
EscherNet will be presented tomorrow at #CVPR. But *now* you can drop a couple of images into our Hugging Face demo to try it out! huggingface.co/spaces/kxic/Es…
Xin Kong@XinKong_IC

Tired of single image to 3D? Check out EscherNet tomorrow @CVPR that can take flexible number of views for 3D generation! THURSDAY, JUNE 20 ORAL: 9:00-10:30, SUMMIT BALLROOM (TOP FLOOR) POSTER: 10:30-12:00, ARCH 4A-E, #69 Try our @Gradio online demo huggingface.co/spaces/kxic/Es…

English
0
0
6
420
Marwan Taher retweetledi
Aalok Patwardhan
Aalok Patwardhan@AalokPat·
From RGB images we can estimate camera rotation, *without* knowledge of camera intrinsics. This also leads to some cool downstream applications - it can complement an IMU .. we call it U-ARE-ME! @AjdDavison @DoC_Rhodes94 @BaeGwangbin
Gwangbin Bae@BaeGwangbin

𝗜𝗠𝗨? How about 𝗨-𝗔𝗥𝗘-𝗠𝗘? In this work, we show how monocular surface normal cues can be used for rotation estimation. callum-rhodes.github.io/U-ARE-ME/ collab w/ @AalokPat, Callum Rhodes, @AjdDavison

English
1
6
42
4.4K
Marwan Taher
Marwan Taher@marwan_ptr·
This allows highly accurate pose estimation even for challenging small and specular objects, which can enable precise manipulation. @ieee_ras_icra (3/3)
English
0
0
4
716
Marwan Taher
Marwan Taher@marwan_ptr·
Using *RGB* images, a NeRF is trained via Instant-NGP. A depth map of the object is rendered to obtain an initial coarse position estimate. The object model is then fitted to the reconstructed density field via a multi-hypothesis iterative optimization scheme. (2/3)
English
1
0
10
1K

Keşfet

@alzugarayign @makezur @XinKong_IC @AjdDavison @eric_dexheimer @_ayoungk @rmurai0610 @_Sir_MuEl