karim416/logdesc

How karim416/logdesc works

LoGDesc is a point-cloud registration system: the documented MVP-RG workflow supplies point-cloud pairs and a pretrained model, then produces a registration result. The source shows LoGDesc registration composing feature extraction, geometric embeddings, transformer matching, and surface features; the README mentions optional keypoint selection. The repository also contains a distinct MVP-RG completion subsystem and registration baselines. Their broader runtime wiring is not fully shown, so those links are omitted where unsupported.

Main entry points: LoGDesc_reg.py LoGDesc Matcher · conditional_transformer.py Conditional Transformer · model_utils.py Completion Utilities · layers.py Layer Builders

Components 23

Point Registration

MVP-RG Pairs MVP_RG/registration/dataset.py
Registration Baselines MVP_RG/registration/models
Pose And Matches

Features And Matching

LoGDesc Matcher models/LoGDesc_reg.py
LoG Features models/LOGDESC.py
Feature Transformer models/lepard_transf.py
Geometric Embedding models/normal_embedding.py
Conditional Transformer models/transformer/conditional_transformer.py
Relative Attention models/transformer/rpe_transformer.py
Vanilla Attention models/transformer/vanilla_transformer.py
Layer Builders models/transformer/layers.py
Surface Features modules/repsurface_utils.py
Polar Utilities modules/polar_recons_utils.py

Point Completion

Completion Dataset MVP_RG/completion/dataset.py
Completion Training MVP_RG/completion/train.py
Completion Models MVP_RG/completion/models
PCN Model MVP_RG/completion/models/pcn.py
Completion Utilities MVP_RG/completion/model_utils.py
Training Utilities MVP_RG/completion/train_utils.py

Point Cloud Operations

Point Cloud Helpers MVP_RG/pcd.py
PointNet Operations MVP_RG/utils/mm3d_pn2
Sparse Convolution MVP_RG/utils/mm3d_pn2/ops/spconv/conv.py

Other components

Researcher

Connections 23

Researcher → MVP-RG Pairs: supplies pairs README.md
Researcher → LoGDesc Matcher: runs matcher README.md
MVP-RG Pairs → Point Cloud Helpers: uses point helpers MVP_RG/registration/dataset.py
LoGDesc Matcher → LoG Features: calls models/LoGDesc_reg.py
LoGDesc Matcher → Geometric Embedding: uses embeddings models/LoGDesc_reg.py
LoGDesc Matcher → Surface Features: constructs surfaces models/LoGDesc_reg.py
LoGDesc Matcher → Conditional Transformer: runs transformer models/LoGDesc_reg.py
LoGDesc Matcher → Pose And Matches: returns matches models/LoGDesc_reg.py
LoG Features → Feature Transformer: uses features models/LOGDESC.py
Conditional Transformer → Relative Attention: uses attention models/transformer/conditional_transformer.py
Conditional Transformer → Vanilla Attention: uses attention models/transformer/conditional_transformer.py
Relative Attention → Layer Builders: builds layers models/transformer/rpe_transformer.py
Vanilla Attention → Layer Builders: builds layers models/transformer/vanilla_transformer.py
Geometric Embedding → Layer Builders: builds dropout models/normal_embedding.py
Surface Features → Polar Utilities: uses utilities modules/repsurface_utils.py
Completion Training → Completion Dataset: loads dataset MVP_RG/completion/train.py
Completion Training → Training Utilities: uses utilities MVP_RG/completion/train.py
Completion Models → PCN Model: includes model MVP_RG/completion/models/vrcnet.py
Completion Models → Completion Utilities: uses utilities MVP_RG/completion/models/vrcnet.py
Completion Utilities → PointNet Operations: uses point operations MVP_RG/completion/model_utils.py
PCN Model → Completion Utilities: uses utilities MVP_RG/completion/models/pcn.py
Registration Baselines → Training Utilities: shares training utilities MVP_RG/registration/models/dcp.py
PointNet Operations → Sparse Convolution: provides operations MVP_RG/utils/mm3d_pn2/ops/spconv/modules.py

Drawn by GitDiagram on November 13, 2026 from the repository's file tree, README and source files. Read as Markdown ·