@makezur profile picture
Kirill Mazur @makezur

PhD student at Dyson Robotics Lab @ Imperial College London

Joined May 2020
  • Tweets 32
  • Following 395
  • Followers 477
10 Photos and videos
Pinned Tweet
Kirill Mazur@makezur
18 Dec 2025
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
0:30
5
24
176
32,664
Kirill Mazur retweeted
Marwan Taher@marwan_ptr
Jun 6
Wanna see a real-time demo of KV-Tracker? Stop by in ExHall F today 11:45 - 13:45! #CVPR2026
0:24
2
12
55
4,735
How to build complete 4D reconstructions from videos? Swing by 4DPM oral presentation today: Mile High Ballroom 3A - 4A, 13:00 #CVPR2026
0:07
3
8
108
10,403
code is released too: github.com/makezur/4D_PM project page: makezur.github.io/4DPM/

GitHub - makezur/4D_PM: [CVPR 2026 Oral] 4D Primitive-Mâché: Glueing Primitives for Persistent 4D...

[CVPR 2026 Oral] 4D Primitive-Mâché: Glueing Primitives for Persistent 4D Scene Reconstruction - makezur/4D_PM

github.com
2
393
4DPM got selected for an oral presentation at #CVPR2026 🤠
Kirill Mazur@makezur
18 Dec 2025
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
0:30
3
5
65
6,952
accepted to #CVPR2026 😎
Kirill Mazur@makezur
18 Dec 2025
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
0:30
1
6
86
7,062
project page: makezur.github.io/4DPM/

4DPM — 4D Primitive-Mâché

Glueing Primitives for Persistent 4D Scene Reconstruction — paper, video and results.

makezur.github.io
137
Kirill Mazur retweeted
Zhenjun Zhao@zhenjun_zhao
20 Dec 2025
4D Primitive-Mâché: Glueing Primitives for Persistent 4D Scene Reconstruction @makezur, @marwan_ptr, @AjdDavison tl;dr: improve SuperPrimitive; scenes->3D primitives->sparse per-primitive SE(3) poses ( 2D correspondences)->per-pixel motion fields arxiv.org/abs/2512.16564
7
33
1,913
Kirill Mazur retweeted
Marwan Taher@marwan_ptr
19 Dec 2025
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
0:18
10
96
704
65,394
Kirill Mazur@makezur
18 Dec 2025
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
0:30
5
24
176
32,664
more replies
Kirill Mazur@makezur
18 Dec 2025
4DPM handles multiple moving entities with distinct trajectories. The result is a complete, replayable 4D reconstruction of all observed scene elements at every timestamp.
0:16
1
1
5
850
Kirill Mazur@makezur
18 Dec 2025
Our method cuts 3D primitives out of the outputs of a feed-forward reconstruction model (π³) and then glues them across time. Each primitive’s motion is represented compactly as a single SE(3) pose, inferred from estimated 2D correspondences via an optimisation pipeline.
0:42
6
602
Kirill Mazur retweeted
Riku Murai@rmurai0610
16 Dec 2024
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)
0:34
42
253
1,434
203,565
Kirill Mazur@makezur
19 Jun 2024
Come by our poster this afternoon! 𝟭𝟳:𝟭𝟱 𝘁𝗼 𝟭𝟴:𝟰𝟱, 𝗔𝗿𝗰𝗵 𝟰𝗔-𝗘
Kirill Mazur@makezur
14 Jun 2024
SuperPrimitives will be presented at #CVPR next week (Wednesday), along with a 𝗿𝗲𝗮𝗹-𝘁𝗶𝗺𝗲 𝗱𝗲𝗺𝗼 on Friday! Our new representation enables dense monocular 3D reconstruction in real-time. No poses required! Project page: makezur.github.io/SuperPrimi…
0:44
3
9
50
5,383
Kirill Mazur@makezur
21 Jun 2024
Demo tomorrow! x.com/makezur/status/1803471…

Kirill Mazur@makezur
19 Jun 2024
Come by our poster this afternoon! 𝟭𝟳:𝟭𝟱 𝘁𝗼 𝟭𝟴:𝟰𝟱, 𝗔𝗿𝗰𝗵 𝟰𝗔-𝗘
1
310
Kirill Mazur@makezur
21 Jun 2024
Curious how DSINE surface normals can be converted into full 3D geometry? Visit our demo of 𝗦𝘂𝗽𝗲𝗿𝗣𝗿𝗶𝗺𝗶𝘁𝗶𝘃𝗲-based real-time 3D reconstruction tomorrow (Friday)! A demo is worth a thousand pictures 🙂 Demo: 10:30 to 18:45, Arch CDE #9 #CVPR #CVPR2024
Kirill Mazur@makezur
14 Jun 2024
SuperPrimitives will be presented at #CVPR next week (Wednesday), along with a 𝗿𝗲𝗮𝗹-𝘁𝗶𝗺𝗲 𝗱𝗲𝗺𝗼 on Friday! Our new representation enables dense monocular 3D reconstruction in real-time. No poses required! Project page: makezur.github.io/SuperPrimi…
0:44
1
14
1,930
Kirill Mazur@makezur
19 Jun 2024
Poster 14!
5
195
Kirill Mazur@makezur
14 Jun 2024
SuperPrimitives will be presented at #CVPR next week (Wednesday), along with a 𝗿𝗲𝗮𝗹-𝘁𝗶𝗺𝗲 𝗱𝗲𝗺𝗼 on Friday! Our new representation enables dense monocular 3D reconstruction in real-time. No poses required! Project page: makezur.github.io/SuperPrimi…
0:44
3
42
208
30,797
Kirill Mazur retweeted
Eric Dexheimer@eric_dexheimer
3 Apr 2024
How can we infer 3D-consistent poses and dense geometry in real-time given only RGB images? 𝗖𝗢𝗠𝗢 decodes dense geometry from a compact and optimizable set of 3D anchor points to enforce 3D consistency. Project page: edexheim.github.io/como/ Work with @AjdDavison 1/n
0:21
3
54
265
41,785
Kirill Mazur@makezur
22 Mar 2024
Code release for SuperPrimitives, and it comes with an interactive GUI! #CVPR2024 github.com/makezur/super_pri… SuperPrimitive is a new 3D representation which enables solving many 3D tasks at the level of image segments.
0:35
1
37
169
22,974
Load more