# meidachen/stpls3d architecture

> Architecture diagram of the GitHub repository meidachen/stpls3d, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-05). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/meidachen/stpls3d
- Repository: https://github.com/meidachen/stpls3d

## Overview

STPLS3D is an annotated aerial point-cloud dataset accompanied by example semantic- and instance-segmentation benchmarks. The user-facing workflow is to prepare/load point clouds, train a segmentation model, then evaluate predictions. The repository contains several largely independent model implementations rather than one shared application pipeline. Edges below are limited to relationships visible in the source index; documented training and evaluation capabilities are represented, but some links to unsampled entry points remain unverified.

Main entry points: [`KPConv-PyTorch/datasets/STPLS3D.py`](https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/datasets/STPLS3D.py) (KPConv loader), [`KPConv-PyTorch/train_STPLS3D.py`](https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/train_STPLS3D.py) (KPConv training), [`HAIS/model/hais/hais.py`](https://github.com/meidachen/stpls3d/blob/main/HAIS/model/hais/hais.py) (HAIS model), [`KPConv-PyTorch/models/architectures.py`](https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/models/architectures.py) (KPConv architecture)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_data["Dataset handling"]
  node_stpls[("STPLS3D data")]
  node_kp_dataset["KPConv loader<br/>[STPLS3D.py]"]
  node_kp_common["Cloud utilities<br/>[common.py]"]
end

subgraph group_semantic["Semantic benchmarks"]
  node_kp_train["KPConv training<br/>[train_STPLS3D.py]"]
  node_kp_model["KPConv architecture<br/>[architectures.py]"]
  node_kp_blocks["KPConv blocks<br/>[blocks.py]"]
  node_kp_kernels["Kernel points<br/>[kernel_points.py]"]
  node_kp_eval["KPConv evaluation<br/>[tester.py]"]
  node_kp_metrics["Segmentation metrics<br/>[metrics.py]"]
  node_randla["RandLA-Net<br/>[RandLANet.py]"]
  node_randla_main["RandLA workflow<br/>[main_STPLS3D.py]"]
  node_scf["SCF-Net<br/>[SCFNet.py]"]
  node_scf_main["SCF workflow<br/>[STPLS3D_main.py]"]
  node_pt_train["Point Transformer training<br/>[train_stpls.py]"]
  node_pt_model["Point Transformer model"]
  node_pt_transform["Point transforms<br/>[transform.py]"]
end

subgraph group_instance["Instance benchmark"]
  node_hais_train["HAIS training<br/>[train.py]"]
  node_hais_data["Instance data loader<br/>[stpls3d_inst.py]"]
  node_hais_model["HAIS model<br/>[hais.py]"]
  node_challenge_eval["Challenge evaluation"]
end

subgraph group_kernels["Point operations"]
  node_hais_ops["HAIS point ops<br/>[hais_ops.py]"]
  node_spconv_modules["Sparse modules<br/>[modules.py]"]
end

node_user(("Researcher"))

node_user -.->|"uses data"| node_stpls
node_stpls -.->|"loads clouds"| node_kp_dataset
node_kp_train -.->|"trains on"| node_kp_dataset
node_kp_train -.->|"trains model"| node_kp_model
node_kp_model -->|"uses blocks"| node_kp_blocks
node_kp_blocks -->|"uses kernels"| node_kp_kernels
node_kp_dataset -->|"uses utilities"| node_kp_common
node_kp_eval -->|"computes metrics"| node_kp_metrics
node_user -.->|"runs benchmark"| node_randla_main
node_randla_main -.->|"trains model"| node_randla
node_user -.->|"runs benchmark"| node_scf_main
node_scf_main -.->|"trains model"| node_scf
node_user -.->|"runs training"| node_pt_train
node_pt_train -->|"trains model"| node_pt_model
node_pt_model -.->|"uses transforms"| node_pt_transform
node_user -.->|"runs training"| node_hais_train
node_hais_train -.->|"loads instances"| node_hais_data
node_hais_model -->|"calls operations"| node_hais_ops
node_hais_model -->|"uses modules"| node_spconv_modules
node_hais_data -->|"uses operations"| node_hais_ops
node_user -.->|"submits results"| node_challenge_eval

click node_kp_dataset "https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/datasets/STPLS3D.py"
click node_kp_common "https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/datasets/common.py"
click node_kp_train "https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/train_STPLS3D.py"
click node_kp_model "https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/models/architectures.py"
click node_kp_blocks "https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/models/blocks.py"
click node_kp_kernels "https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/kernels/kernel_points.py"
click node_kp_eval "https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/utils/tester.py"
click node_kp_metrics "https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/utils/metrics.py"
click node_randla "https://github.com/meidachen/stpls3d/blob/main/RandLA-Net/RandLANet.py"
click node_randla_main "https://github.com/meidachen/stpls3d/blob/main/RandLA-Net/main_STPLS3D.py"
click node_scf "https://github.com/meidachen/stpls3d/blob/main/SCF-Net/SCFNet.py"
click node_scf_main "https://github.com/meidachen/stpls3d/blob/main/SCF-Net/STPLS3D_main.py"
click node_pt_train "https://github.com/meidachen/stpls3d/blob/main/point-transformer/tool/train_stpls.py"
click node_pt_model "https://github.com/meidachen/stpls3d/blob/main/point-transformer/model/pointtransformer/pointtransformer_seg.py"
click node_pt_transform "https://github.com/meidachen/stpls3d/blob/main/point-transformer/util/transform.py"
click node_hais_train "https://github.com/meidachen/stpls3d/blob/main/HAIS/train.py"
click node_hais_data "https://github.com/meidachen/stpls3d/blob/main/HAIS/data/stpls3d_inst.py"
click node_hais_model "https://github.com/meidachen/stpls3d/blob/main/HAIS/model/hais/hais.py"
click node_hais_ops "https://github.com/meidachen/stpls3d/blob/main/HAIS/lib/hais_ops/functions/hais_ops.py"
click node_spconv_modules "https://github.com/meidachen/stpls3d/blob/main/HAIS/lib/spconv/spconv/modules.py"
click node_challenge_eval "https://github.com/meidachen/stpls3d/blob/main/HAIS/STPLS3DInstanceSegmentationChallenge_Codalab_Evaluate.py"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_stpls,node_kp_dataset,node_kp_common toneBlue
class node_kp_train,node_kp_model,node_kp_blocks,node_kp_kernels,node_kp_eval,node_kp_metrics,node_randla,node_randla_main,node_scf,node_scf_main,node_pt_train,node_pt_model,node_pt_transform toneAmber
class node_hais_train,node_hais_data,node_hais_model,node_challenge_eval toneMint
class node_hais_ops,node_spconv_modules toneRose
class node_user toneIndigo
```

## Components

### Dataset handling

- **STPLS3D data**
- [KPConv loader](https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/datasets/STPLS3D.py): `KPConv-PyTorch/datasets/STPLS3D.py`
- [Cloud utilities](https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/datasets/common.py): `KPConv-PyTorch/datasets/common.py`

### Semantic benchmarks

- [KPConv training](https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/train_STPLS3D.py): `KPConv-PyTorch/train_STPLS3D.py`
- [KPConv architecture](https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/models/architectures.py): `KPConv-PyTorch/models/architectures.py`
- [KPConv blocks](https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/models/blocks.py): `KPConv-PyTorch/models/blocks.py`
- [Kernel points](https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/kernels/kernel_points.py): `KPConv-PyTorch/kernels/kernel_points.py`
- [KPConv evaluation](https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/utils/tester.py): `KPConv-PyTorch/utils/tester.py`
- [Segmentation metrics](https://github.com/meidachen/stpls3d/blob/main/KPConv-PyTorch/utils/metrics.py): `KPConv-PyTorch/utils/metrics.py`
- [RandLA-Net](https://github.com/meidachen/stpls3d/blob/main/RandLA-Net/RandLANet.py): `RandLA-Net/RandLANet.py`
- [RandLA workflow](https://github.com/meidachen/stpls3d/blob/main/RandLA-Net/main_STPLS3D.py): `RandLA-Net/main_STPLS3D.py`
- [SCF-Net](https://github.com/meidachen/stpls3d/blob/main/SCF-Net/SCFNet.py): `SCF-Net/SCFNet.py`
- [SCF workflow](https://github.com/meidachen/stpls3d/blob/main/SCF-Net/STPLS3D_main.py): `SCF-Net/STPLS3D_main.py`
- [Point Transformer training](https://github.com/meidachen/stpls3d/blob/main/point-transformer/tool/train_stpls.py): `point-transformer/tool/train_stpls.py`
- [Point Transformer model](https://github.com/meidachen/stpls3d/blob/main/point-transformer/model/pointtransformer/pointtransformer_seg.py): `point-transformer/model/pointtransformer/pointtransformer_seg.py`
- [Point transforms](https://github.com/meidachen/stpls3d/blob/main/point-transformer/util/transform.py): `point-transformer/util/transform.py`

### Instance benchmark

- [HAIS training](https://github.com/meidachen/stpls3d/blob/main/HAIS/train.py): `HAIS/train.py`
- [Instance data loader](https://github.com/meidachen/stpls3d/blob/main/HAIS/data/stpls3d_inst.py): `HAIS/data/stpls3d_inst.py`
- [HAIS model](https://github.com/meidachen/stpls3d/blob/main/HAIS/model/hais/hais.py): `HAIS/model/hais/hais.py`
- [Challenge evaluation](https://github.com/meidachen/stpls3d/blob/main/HAIS/STPLS3DInstanceSegmentationChallenge_Codalab_Evaluate.py): `HAIS/STPLS3DInstanceSegmentationChallenge_Codalab_Evaluate.py`

### Point operations

- [HAIS point ops](https://github.com/meidachen/stpls3d/blob/main/HAIS/lib/hais_ops/functions/hais_ops.py): `HAIS/lib/hais_ops/functions/hais_ops.py`
- [Sparse modules](https://github.com/meidachen/stpls3d/blob/main/HAIS/lib/spconv/spconv/modules.py): `HAIS/lib/spconv/spconv/modules.py`

### Other components

- **Researcher**

## Connections

- Researcher → STPLS3D data: uses data
- STPLS3D data → KPConv loader: loads clouds
- KPConv training → KPConv loader: trains on
- KPConv training → KPConv architecture: trains model
- KPConv architecture → KPConv blocks: uses blocks (evidence: `KPConv-PyTorch/models/blocks.py`)
- KPConv blocks → Kernel points: uses kernels (evidence: `KPConv-PyTorch/models/blocks.py`)
- KPConv loader → Cloud utilities: uses utilities (evidence: `KPConv-PyTorch/datasets/STPLS3D.py`)
- KPConv evaluation → Segmentation metrics: computes metrics (evidence: `KPConv-PyTorch/utils/tester.py`)
- Researcher → RandLA workflow: runs benchmark
- RandLA workflow → RandLA-Net: trains model
- Researcher → SCF workflow: runs benchmark
- SCF workflow → SCF-Net: trains model
- Researcher → Point Transformer training: runs training
- Point Transformer training → Point Transformer model: trains model (evidence: `point-transformer/tool/train_stpls.py`)
- Point Transformer model → Point transforms: uses transforms
- Researcher → HAIS training: runs training
- HAIS training → Instance data loader: loads instances
- HAIS model → HAIS point ops: calls operations (evidence: `HAIS/model/hais/hais.py`)
- HAIS model → Sparse modules: uses modules (evidence: `HAIS/model/hais/hais.py`)
- Instance data loader → HAIS point ops: uses operations (evidence: `HAIS/data/stpls3d_inst.py`)
- Researcher → Challenge evaluation: submits results
