PhD Candidate in Computing @imperialcollege
London, England
Joined May 2026
- Tweets14
- Following110
- Followers107
- Likes18
Nicholas Fry retweeted
AgentSTAR is an agentic method for both reconstruction and tracking of articulated objects from monocular videos.
We can now track through rapid motion, severe occlusion and thin objects!
There’s no spoon but we can still track it 🥄
I'll be presenting our work at #ECCV2026 in Malmö! 🇸🇪 Find me at the poster and demo sessions:
Poster
📅 Thu 10 Sep, 10:30
📍 ExHall #272
Demo
📅 Sat 12 Sep, 8:00
📍 ExHall
Also at NeuSLAM Workshop 🤖:
Tue 8 Sep, 14:00, Malmö Arena Hotel
Nicholas Fry retweeted
(1/4)
Two (co-)first-author papers at #ECCV2026 in Malmö 🇸🇪
📅 Thu 10 Sep, 10:30
📍 PixVOD → ExHall #78
📍 MLP-Splatting → ExHall #277
Code is out !
nmjfry.github.io/triangle-sp…
There is also an experimental branch using the rasteriser from kenji-tojo.github.io/publica… by @tojo_kenji - I haven't properly implemented densification, but it already looks promising.
What if we had converged to using different hardware architectures for visual computing ?
Check out how we render 3D Gaussians on a Coarse-Grained Reconfigurable Array (CGRA): nmjfry.github.io/ipu-3dgs/
Presented at EGSR 2026.
Work with @alzugarayign, @MarkPupilli, @paulhjkelly and @AjdDavison
Nicholas Fry retweeted
A brilliant talk from @deepfry_n from EGSR 2026 on his paper Rendering 3D Gaussians on a Graph Processor. We assign Gaussians to tiles, fully with on-chip SRAM, and pass them around efficiently as needed as the virtual camera moves viewpoint. Joint work with @alzugarayign, @MarkPupilli and Paul Kelly. Full talk here: youtu.be/OHfQk9J-CAg?si=1DyP…
Nicholas Fry retweeted
What if you could turn any number of photos (3, 8, 15, or even 60) into one clean 3D surface (pts & mesh) with Flow Matching?
Check out our new work, Surflo: Consistent 3D Surface Flow Model with Global State. 🧵
1/n
🔗anttwo.github.io/surflo/
Nicholas Fry retweeted
(1/5) Excited to introduce MLP-Splatting, a novel way to decompose a scene into object-/part-level primitives!
Each object can be represented with a few tens of light-field-based primitives (c.f. 3DGS: hundreds+).
w/ @cheng_yuzh21715 @XinKong_IC @paulhjkelly @AjdDavison
(1/3) I am happy to share our work on Triangle Splatting SLAM!
We show the first RGB-D SLAM to use differentiable triangles as a 3D map representation.
Our method enables online mesh-based deformations and collision checking via on-the-fly Delaunay triangulation.
(2/3) Work with @eric_dexheimer, @makezur, @paulhjkelly, @AjdDavison
nmjfry.github.io/triangle-sp…
Here are some fun demos!
Nicholas Fry retweeted
Join us on 3 Jun. at 8 AM in Room 4EF for the On-Sensor Vision Workshop at @CVPR! #CVPR2026
onsensor-vision.github.io/
Through invited talks and posters, we’ll explore the future of on-sensor vision: low-power/latency edge computer vision enabled by sensor-side computation.
