Skip to content

KFS⁴LAM Reconstruction Parameters

Latest

Choose a tag to compare

@XinweiJu XinweiJu released this 15 Oct 12:01
· 2 commits to code since this release

KFS⁴LAM Reconstruction Parameters

25 reconstruction archives covering 14 C3VD and 11 C3VD-v2 sequences. Each includes Gaussian parameters, temporal deformation coefficients, camera poses, intrinsics, keyframe indices and a portable configuration.

Download and extract an archive, then use render_checkpoint.py --backbone splatam --checkpoint params.npz --frames all --output rendered from the code repository. The splatam option selects the implementation backend; the method is KFS⁴LAM. Archive-internal directory names retain the backend identifier for compatibility. Rendering saved poses does not require the source dataset or retraining.

Verify downloads with sha256sum -c SHA256SUMS. See manifest.json and metrics.json for per-sequence metadata and results. The original RGB-D and fold maps are required for a new SLAM run; RGB-D is required for evaluation against ground truth.