explaingit

makezur/4d_pm

14PythonAudience · researcherComplexity · 5/5Setup · hard

TLDR

CVPR 2026 research code that reconstructs 4D scenes from video by fitting geometric primitives to objects and tracking them persistently over time, even when they temporarily leave the frame.

Mindmap

mindmap
  root((4d_pm))
    What it does
      4D scene reconstruction
      Object tracking over time
      Persistent object presence
    Pipeline
      Frontend geometry
      Pi3 segmentation
      SAM 2 masking
      Gauss-Newton optimizer
    Output
      Geometric primitives
      4D replay
      Time-extended scene
    Demos
      Robot arm dataset
      Drawer scenario
      Fridge scenario
    Requirements
      NVIDIA GPU
      CUDA
      PyTorch
Click or tap to explore — scroll the page freely

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

Things people build with this

USE CASE 1

Run the included drawer or fridge demo to reconstruct how objects move when they temporarily leave the frame.

USE CASE 2

Use the codebase as a baseline for CVPR 2026 comparisons in 4D scene reconstruction.

USE CASE 3

Apply the Gauss-Newton primitive fitting pipeline to a custom video dataset.

Tech stack

PythonPyTorchCUDASAM 2Pi3Rerun

Getting it running

Difficulty · hard Time to first run · 1day+

Requires an NVIDIA GPU with CUDA and model checkpoints downloaded via the included install script.

License not specified in the explanation.

In plain English

4D Primitive-Mache is a research codebase from a paper accepted as an oral presentation at CVPR 2026, a major academic conference in computer vision. The project addresses a problem called persistent 4D scene reconstruction, which means building a model of a physical scene that tracks how objects move and change over time across multiple video frames, not just capturing a single static snapshot. The core idea is to represent scenes using geometric primitives: simple shapes like ellipsoids or similar building blocks that can be positioned, oriented, and deformed. The paper proposes a method for fitting these primitives to video footage and tracking them persistently over time, even when objects temporarily disappear from view (for example, when a drawer closes and hides its contents). This property, called object permanence in the demo configurations, is what the word "persistent" in the title refers to. The codebase is organized into three main parts. The frontend handles geometry estimation and object segmentation, using two external models called Pi3 and SAM 2. The core module handles mathematical optimization: specifically a Gauss-Newton solver that fits the primitives to the observed data. The object mapper handles motion tracking and assembles what the authors call 4D replay, which is a time-extended representation of the scene that can be replayed or inspected after reconstruction. Running the system requires an NVIDIA graphics card, CUDA, and the PyTorch deep learning library. An install script sets up the environment, downloads model checkpoints, and configures paths automatically. Demo configurations for a robot arm dataset and two object-permanence scenarios (a drawer and a fridge) are included. The README is technical and assumes familiarity with computer vision research. There is no graphical interface, results are visualized using an external tool called Rerun.

Copy-paste prompts

Prompt 1
I want to run the 4d_pm drawer demo. Walk me through running the install script, downloading the checkpoints, and executing the demo configuration step by step.
Prompt 2
I am comparing 4D scene reconstruction methods for a paper. How does 4d_pm represent objects differently from NeRF-based approaches, and what are the tradeoffs for tracking occluded objects?
Prompt 3
Show me how to replace the SAM 2 segmentation frontend in 4d_pm with a custom segmentation model while keeping the Gauss-Newton optimizer unchanged.
Open on GitHub → Explain another repo

← makezur on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.