# vidhigadge/steelvision-dl architecture

> Architecture diagram of the GitHub repository vidhigadge/steelvision-dl, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-04). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/vidhigadge/steelvision-dl
- Repository: https://github.com/vidhigadge/steelvision-dl

## Overview

SteelVision is a steel-defect research prototype with a user-facing inspection workflow and supporting model-development experiments. In standard analysis, a user submits an image through the interface; the API classifies it, reports confidence and entropy, and can request Grad-CAM and localization with mapped defect boxes. A separate localization experiment varies input resolution. The graph also preserves dataset, training, super-resolution, reliability, and deployment-optimization capabilities documented in the README. Source excerpts establish API-to-service and training-to-model calls; unsampled UI wiring and several research links are shown only as unverified where appropriate.

Main entry points: [`src/models/resnet50_finetune.py`](https://github.com/vidhigadge/steelvision-dl/blob/main/src/models/resnet50_finetune.py) (Fine-Tuned ResNet), [`src/training`](https://github.com/vidhigadge/steelvision-dl/tree/main/src/training) (Model Training), [`src/api/main.py`](https://github.com/vidhigadge/steelvision-dl/blob/main/src/api/main.py) (FastAPI Endpoints), [`src/evaluation`](https://github.com/vidhigadge/steelvision-dl/tree/main/src/evaluation) (Model Evaluation)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_client["User Interface"]
  node_ui["Streamlit Interface<br/>[app.py]"]
end

subgraph group_serve["Inspection Services"]
  node_api["FastAPI Endpoints<br/>[main.py]"]
  node_classifier["ONNX Classifier<br/>[main.py]"]
  node_gradcam["Grad-CAM Service<br/>[gradcam_service.py]"]
  node_localize["YOLO Localization"]
  node_mapping["Box Coordinate Mapping"]
end

subgraph group_research["Model Research"]
  node_cnn["Baseline CNN<br/>[cnn.py]"]
  node_resnet["ResNet-50<br/>[resnet50.py]"]
  node_finetune["Fine-Tuned ResNet"]
  node_training["Model Training"]
  node_evaluation["Model Evaluation"]
  node_resolution["Resolution Experiments"]
  node_superres["Super-Resolution Study"]
  node_reliability["Uncertainty and Calibration"]
  node_shap["SHAP Explanations<br/>[shap_explain.py]"]
end

subgraph group_data["Dataset Workflow"]
  node_datafiles[("Image Splits")]
  node_dataops["Dataset Preparation"]
  node_loaders["Classification Loaders<br/>[loaders.py]"]
end

subgraph group_deploy["Deployment Optimization"]
  node_optimization["ONNX Optimization"]
end

node_user(("Inspector"))

node_user -.->|"submits image"| node_ui
node_ui -.->|"requests analysis"| node_api
node_api -->|"runs inference"| node_classifier
node_api -->|"requests explanation"| node_gradcam
node_api -->|"requests localization"| node_localize
node_localize -->|"maps boxes"| node_mapping
node_training -->|"loads batches"| node_loaders
node_training -->|"trains model"| node_finetune
node_training -->|"trains baseline"| node_cnn
node_training -->|"trains transfer model"| node_resnet
node_evaluation -->|"loads evaluation data"| node_loaders
node_evaluation -->|"evaluates model"| node_finetune
node_resolution -->|"measures accuracy"| node_finetune
node_superres -.->|"compares classification"| node_finetune
node_reliability -->|"analyzes predictions"| node_finetune
node_shap -->|"explains predictions"| node_finetune
node_optimization -.->|"optimizes inference"| node_finetune
node_dataops -.->|"prepares splits"| node_datafiles

click node_ui "https://github.com/vidhigadge/steelvision-dl/blob/main/src/ui/app.py"
click node_api "https://github.com/vidhigadge/steelvision-dl/blob/main/src/api/main.py"
click node_classifier "https://github.com/vidhigadge/steelvision-dl/blob/main/src/api/main.py"
click node_gradcam "https://github.com/vidhigadge/steelvision-dl/blob/main/src/explainability/gradcam_service.py"
click node_localize "https://github.com/vidhigadge/steelvision-dl/blob/main/src/detection/localization_service.py"
click node_mapping "https://github.com/vidhigadge/steelvision-dl/blob/main/src/detection/coordinate_mapping.py"
click node_datafiles "https://github.com/vidhigadge/steelvision-dl/tree/main/data/splits"
click node_dataops "https://github.com/vidhigadge/steelvision-dl/tree/main/src/data"
click node_loaders "https://github.com/vidhigadge/steelvision-dl/blob/main/src/data/loaders.py"
click node_cnn "https://github.com/vidhigadge/steelvision-dl/blob/main/src/models/cnn.py"
click node_resnet "https://github.com/vidhigadge/steelvision-dl/blob/main/src/models/resnet50.py"
click node_finetune "https://github.com/vidhigadge/steelvision-dl/blob/main/src/models/resnet50_finetune.py"
click node_training "https://github.com/vidhigadge/steelvision-dl/tree/main/src/training"
click node_evaluation "https://github.com/vidhigadge/steelvision-dl/tree/main/src/evaluation"
click node_resolution "https://github.com/vidhigadge/steelvision-dl/blob/main/src/evaluation/evaluate_resolution.py"
click node_superres "https://github.com/vidhigadge/steelvision-dl/tree/main/src/super_resolution"
click node_reliability "https://github.com/vidhigadge/steelvision-dl/tree/main/src/reliability"
click node_shap "https://github.com/vidhigadge/steelvision-dl/blob/main/src/explainability/shap_explain.py"
click node_optimization "https://github.com/vidhigadge/steelvision-dl/tree/main/src/optimization"

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_ui toneBlue
class node_api,node_classifier,node_gradcam,node_localize,node_mapping toneAmber
class node_cnn,node_resnet,node_finetune,node_training,node_evaluation,node_resolution,node_superres,node_reliability,node_shap toneMint
class node_datafiles,node_dataops,node_loaders toneRose
class node_optimization,node_user toneIndigo
```

## Components

### User Interface

- [Streamlit Interface](https://github.com/vidhigadge/steelvision-dl/blob/main/src/ui/app.py): `src/ui/app.py`

### Inspection Services

- [FastAPI Endpoints](https://github.com/vidhigadge/steelvision-dl/blob/main/src/api/main.py): `src/api/main.py`
- [ONNX Classifier](https://github.com/vidhigadge/steelvision-dl/blob/main/src/api/main.py): `src/api/main.py`
- [Grad-CAM Service](https://github.com/vidhigadge/steelvision-dl/blob/main/src/explainability/gradcam_service.py): `src/explainability/gradcam_service.py`
- [YOLO Localization](https://github.com/vidhigadge/steelvision-dl/blob/main/src/detection/localization_service.py): `src/detection/localization_service.py`
- [Box Coordinate Mapping](https://github.com/vidhigadge/steelvision-dl/blob/main/src/detection/coordinate_mapping.py): `src/detection/coordinate_mapping.py`

### Model Research

- [Baseline CNN](https://github.com/vidhigadge/steelvision-dl/blob/main/src/models/cnn.py): `src/models/cnn.py`
- [ResNet-50](https://github.com/vidhigadge/steelvision-dl/blob/main/src/models/resnet50.py): `src/models/resnet50.py`
- [Fine-Tuned ResNet](https://github.com/vidhigadge/steelvision-dl/blob/main/src/models/resnet50_finetune.py): `src/models/resnet50_finetune.py`
- [Model Training](https://github.com/vidhigadge/steelvision-dl/tree/main/src/training): `src/training`
- [Model Evaluation](https://github.com/vidhigadge/steelvision-dl/tree/main/src/evaluation): `src/evaluation`
- [Resolution Experiments](https://github.com/vidhigadge/steelvision-dl/blob/main/src/evaluation/evaluate_resolution.py): `src/evaluation/evaluate_resolution.py`
- [Super-Resolution Study](https://github.com/vidhigadge/steelvision-dl/tree/main/src/super_resolution): `src/super_resolution`
- [Uncertainty and Calibration](https://github.com/vidhigadge/steelvision-dl/tree/main/src/reliability): `src/reliability`
- [SHAP Explanations](https://github.com/vidhigadge/steelvision-dl/blob/main/src/explainability/shap_explain.py): `src/explainability/shap_explain.py`

### Dataset Workflow

- [Image Splits](https://github.com/vidhigadge/steelvision-dl/tree/main/data/splits): `data/splits`
- [Dataset Preparation](https://github.com/vidhigadge/steelvision-dl/tree/main/src/data): `src/data`
- [Classification Loaders](https://github.com/vidhigadge/steelvision-dl/blob/main/src/data/loaders.py): `src/data/loaders.py`

### Deployment Optimization

- [ONNX Optimization](https://github.com/vidhigadge/steelvision-dl/tree/main/src/optimization): `src/optimization`

### Other components

- **Inspector**

## Connections

- Inspector → Streamlit Interface: submits image
- Streamlit Interface → FastAPI Endpoints: requests analysis
- FastAPI Endpoints → ONNX Classifier: runs inference (evidence: `src/api/main.py`)
- FastAPI Endpoints → Grad-CAM Service: requests explanation (evidence: `src/api/main.py`)
- FastAPI Endpoints → YOLO Localization: requests localization (evidence: `src/api/main.py`)
- YOLO Localization → Box Coordinate Mapping: maps boxes (evidence: `src/detection/localization_service.py`)
- Model Training → Classification Loaders: loads batches (evidence: `src/training/train_resnet_finetune.py`)
- Model Training → Fine-Tuned ResNet: trains model (evidence: `src/training/train_resnet_finetune.py`)
- Model Training → Baseline CNN: trains baseline (evidence: `src/training/train_cnn.py`)
- Model Training → ResNet-50: trains transfer model (evidence: `src/training/train_resnet.py`)
- Model Evaluation → Classification Loaders: loads evaluation data (evidence: `src/evaluation/evaluate_resnet_finetune.py`)
- Model Evaluation → Fine-Tuned ResNet: evaluates model (evidence: `src/evaluation/evaluate_resnet_finetune.py`)
- Resolution Experiments → Fine-Tuned ResNet: measures accuracy (evidence: `src/evaluation/evaluate_resolution.py`)
- Super-Resolution Study → Fine-Tuned ResNet: compares classification (evidence: `README.md`)
- Uncertainty and Calibration → Fine-Tuned ResNet: analyzes predictions (evidence: `src/reliability/uncertainty.py`)
- SHAP Explanations → Fine-Tuned ResNet: explains predictions (evidence: `src/explainability/shap_explain.py`)
- ONNX Optimization → Fine-Tuned ResNet: optimizes inference (evidence: `README.md`)
- Dataset Preparation → Image Splits: prepares splits
