# proteus1991/griddehazenet architecture

> Architecture diagram of the GitHub repository proteus1991/griddehazenet, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-07). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/proteus1991/griddehazenet
- Repository: https://github.com/proteus1991/griddehazenet

## Overview

GridDehazeNet is a PyTorch image-dehazing project with training and SOTS evaluation workflows for indoor and outdoor images. Testing takes a hazy image and its clear reference from the selected validation dataset, runs the network, measures PSNR/SSIM, and can save dehazed results. Training loads paired crops, optimizes the network with smooth-L1 and perceptual losses, validates, and saves weights. The README identifies test.py as the testing entry point, but it was not excerpted; its internal calls are therefore not asserted.

Main entry points: [`train.py`](https://github.com/proteus1991/griddehazenet/blob/master/train.py) (Training entry), [`test.py`](https://github.com/proteus1991/griddehazenet/blob/master/test.py) (Testing entry), [`utils.py`](https://github.com/proteus1991/griddehazenet/blob/master/utils.py) (Metrics and validation), [`val_data.py`](https://github.com/proteus1991/griddehazenet/blob/master/val_data.py) (Validation pairs)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_workflow["Train and evaluate"]
  node_train_entry["Training entry<br/>[train.py]"]
  node_test_entry["Testing entry<br/>[test.py]"]
end

subgraph group_data["Image datasets"]
  node_train_data["Training pairs<br/>[train_data.py]"]
  node_train_images["Train images"]
  node_val_data["Validation pairs<br/>[val_data.py]"]
  node_sots_images["SOTS images"]
end

subgraph group_network["Dehazing network"]
  node_grid_model["GridDehazeNet<br/>[model.py]"]
  node_rdb["Residual dense block"]
end

subgraph group_evaluation["Loss and metrics"]
  node_perceptual["Perceptual loss<br/>[perceptual.py]"]
  node_metrics["Metrics and validation<br/>[utils.py]"]
end

subgraph group_artifacts["Model artifacts"]
  node_weights["Indoor/outdoor weights"]
  node_results["Dehazed results"]
  node_logs["Training logs"]
end

node_user(("Researcher"))

node_user -->|"starts training"| node_train_entry
node_user -->|"starts testing"| node_test_entry
node_train_entry -->|"loads pairs"| node_train_data
node_train_data -->|"reads images"| node_train_images
node_train_entry -->|"loads validation"| node_val_data
node_val_data -->|"reads pairs"| node_sots_images
node_train_entry -->|"trains"| node_grid_model
node_grid_model -->|"uses blocks"| node_rdb
node_train_entry -->|"calculates loss"| node_perceptual
node_train_entry -->|"validates"| node_metrics
node_metrics -->|"saves images"| node_results
node_metrics -->|"writes logs"| node_logs
node_train_entry -->|"loads and saves"| node_weights
node_test_entry -.->|"uses validation data"| node_val_data
node_test_entry -.->|"runs inference"| node_grid_model
node_test_entry -.->|"reports metrics"| node_metrics
node_test_entry -.->|"loads weights"| node_weights
node_test_entry -.->|"saves images"| node_results

click node_train_entry "https://github.com/proteus1991/griddehazenet/blob/master/train.py"
click node_test_entry "https://github.com/proteus1991/griddehazenet/blob/master/test.py"
click node_train_data "https://github.com/proteus1991/griddehazenet/blob/master/train_data.py"
click node_train_images "https://github.com/proteus1991/griddehazenet/tree/master/data/train"
click node_val_data "https://github.com/proteus1991/griddehazenet/blob/master/val_data.py"
click node_sots_images "https://github.com/proteus1991/griddehazenet/tree/master/data/test/SOTS"
click node_grid_model "https://github.com/proteus1991/griddehazenet/blob/master/model.py"
click node_rdb "https://github.com/proteus1991/griddehazenet/blob/master/residual_dense_block.py"
click node_perceptual "https://github.com/proteus1991/griddehazenet/blob/master/perceptual.py"
click node_metrics "https://github.com/proteus1991/griddehazenet/blob/master/utils.py"
click node_weights "https://github.com/proteus1991/griddehazenet/blob/master/indoor_haze_best_3_6"
click node_results "https://github.com/proteus1991/griddehazenet/tree/master/indoor_results"
click node_logs "https://github.com/proteus1991/griddehazenet/tree/master/training_log"

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_train_entry,node_test_entry toneBlue
class node_train_data,node_train_images,node_val_data,node_sots_images toneAmber
class node_grid_model,node_rdb toneMint
class node_perceptual,node_metrics toneRose
class node_weights,node_results,node_logs,node_user toneIndigo
```

## Components

### Train and evaluate

- [Training entry](https://github.com/proteus1991/griddehazenet/blob/master/train.py): `train.py`
- [Testing entry](https://github.com/proteus1991/griddehazenet/blob/master/test.py): `test.py`

### Image datasets

- [Training pairs](https://github.com/proteus1991/griddehazenet/blob/master/train_data.py): `train_data.py`
- [Train images](https://github.com/proteus1991/griddehazenet/tree/master/data/train): `data/train`
- [Validation pairs](https://github.com/proteus1991/griddehazenet/blob/master/val_data.py): `val_data.py`
- [SOTS images](https://github.com/proteus1991/griddehazenet/tree/master/data/test/SOTS): `data/test/SOTS`

### Dehazing network

- [GridDehazeNet](https://github.com/proteus1991/griddehazenet/blob/master/model.py): `model.py`
- [Residual dense block](https://github.com/proteus1991/griddehazenet/blob/master/residual_dense_block.py): `residual_dense_block.py`

### Loss and metrics

- [Perceptual loss](https://github.com/proteus1991/griddehazenet/blob/master/perceptual.py): `perceptual.py`
- [Metrics and validation](https://github.com/proteus1991/griddehazenet/blob/master/utils.py): `utils.py`

### Model artifacts

- [Indoor/outdoor weights](https://github.com/proteus1991/griddehazenet/blob/master/indoor_haze_best_3_6): `indoor_haze_best_3_6`
- [Dehazed results](https://github.com/proteus1991/griddehazenet/tree/master/indoor_results): `indoor_results`
- [Training logs](https://github.com/proteus1991/griddehazenet/tree/master/training_log): `training_log`

### Other components

- **Researcher**

## Connections

- Researcher → Training entry: starts training (evidence: `README.md`)
- Researcher → Testing entry: starts testing (evidence: `README.md`)
- Training entry → Training pairs: loads pairs (evidence: `train.py`)
- Training pairs → Train images: reads images (evidence: `train_data.py`)
- Training entry → Validation pairs: loads validation (evidence: `train.py`)
- Validation pairs → SOTS images: reads pairs (evidence: `val_data.py`)
- Training entry → GridDehazeNet: trains (evidence: `train.py`)
- GridDehazeNet → Residual dense block: uses blocks (evidence: `model.py`)
- Training entry → Perceptual loss: calculates loss (evidence: `train.py`)
- Training entry → Metrics and validation: validates (evidence: `train.py`)
- Metrics and validation → Dehazed results: saves images (evidence: `utils.py`)
- Metrics and validation → Training logs: writes logs (evidence: `utils.py`)
- Training entry → Indoor/outdoor weights: loads and saves (evidence: `train.py`)
- Testing entry → Validation pairs: uses validation data (evidence: `README.md`)
- Testing entry → GridDehazeNet: runs inference (evidence: `README.md`)
- Testing entry → Metrics and validation: reports metrics (evidence: `README.md`)
- Testing entry → Indoor/outdoor weights: loads weights (evidence: `README.md`)
- Testing entry → Dehazed results: saves images (evidence: `README.md`)
