GARfVDB: Cluster visualization enhancements

[swahtz]

Moved clustering and filtering logic into a module
Added GaussianSplat3d filtering for outlier positions, opacities and scales
Added filtering of spatially incoherent clusters

[Copilot]

Pull request overview

This PR enhances the GARfVDB segmentation visualization capabilities by introducing a new clustering and filtering module for instance segmentation, along with a dedicated cluster visualization script. The changes move clustering logic into a reusable module and add filtering capabilities for outlier removal and spatial coherence validation.

Changes:

• Added a new clustering module (garfvdb/evaluation/clustering) with PCA, UMAP, and HDBSCAN-based clustering pipeline
• Created visualize_segmentation_clusters.py to visualize pre-computed segmentation clusters with spatial variance filtering
• Refactored encoder grid sampling in model.py into a reusable method and added get_gaussian_affinity_output for per-Gaussian feature extraction
• Added world_to_camera parameter to view_checkpoint.py rendering pipeline
• Updated dependencies to include cuml, cupy, and matplotlib

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 8 comments.

Show a summary per file

File	Description
visualize_segmentation_clusters.py	New script for visualizing segmentation clusters with filtering options based on spatial variance
garfvdb/evaluation/clustering/clustering.py	New module implementing GPU-accelerated clustering pipeline (PCA → UMAP → HDBSCAN)
garfvdb/evaluation/clustering/init.py	Module initialization exposing clustering functions
garfvdb/model.py	Refactored encoder grid sampling into _sample_encoder_grids_at_gaussians() and added get_gaussian_affinity_output() for per-Gaussian feature extraction
view_checkpoint.py	Added world_to_camera parameter to rendering pipeline and reorganized imports
pyproject.toml	Added cuml dependency for GPU-accelerated machine learning
garfvdb_environment.yml	Added rapidsai cuml, cupy, and matplotlib to conda environment

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.