# deeptanu100/bone-fracture-crack-detection architecture

> Architecture diagram of the GitHub repository deeptanu100/bone-fracture-crack-detection, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-11-22). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/deeptanu100/bone-fracture-crack-detection
- Repository: https://github.com/deeptanu100/bone-fracture-crack-detection

## Overview

This repository develops and deploys a local radiograph fracture classifier. The user workflow is to submit an X-ray through the Gradio app or CLI, run ONNX inference, and receive a threshold-based fracture, normal, or human-review result with an overlay. The training lifecycle loads BoneFract data, trains a compact model, evaluates it on locked and C17 subtype benchmarks, and exports FP32/INT8 ONNX artifacts. Runtime and training paths are documented or visible in excerpts; some subsystem details are represented at the directory level because their source was not sampled.

Main entry points: [`src/train.py`](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/train.py) (Training loop), [`src/evaluate.py`](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/evaluate.py) (Test evaluation), [`src/model.py`](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/model.py) (Fracture classifier), [`src/infer.py`](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/infer.py) (Inference and overlays)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_runtime["Local diagnosis"]
  node_app["Gradio application<br/>[app.py]"]
  node_inference["Inference and overlays<br/>[infer.py]"]
  node_onnxmodel["INT8 ONNX model"]
end

subgraph group_training["Model development"]
  node_bonefract["BoneFract loader<br/>[dataset.py]"]
  node_c17["C17 subtype loader<br/>[dataset.py]"]
  node_transforms["Radiograph transforms<br/>[dataset.py]"]
  node_trainer["Training loop<br/>[train.py]"]
  node_network["Fracture classifier<br/>[model.py]"]
  node_gradcam["Grad-CAM hooks<br/>[model.py]"]
  node_config["Pipeline configuration<br/>[config.yaml]"]
end

subgraph group_evaluation["Validation and export"]
  node_evaluator["Test evaluation<br/>[evaluate.py]"]
  node_exporter["ONNX export and quantization<br/>[export_onnx.py]"]
end

node_patient(("Image submitter"))
node_cli(("CLI inference user"))

node_patient -->|"submits image"| node_app
node_app -->|"calls prediction"| node_inference
node_cli -->|"runs inference"| node_inference
node_inference -->|"loads model"| node_onnxmodel
node_trainer -->|"loads training data"| node_bonefract
node_trainer -->|"uses transforms"| node_transforms
node_trainer -->|"trains model"| node_network
node_trainer -->|"reads settings"| node_config
node_evaluator -->|"loads test data"| node_bonefract
node_evaluator -->|"benchmarks subtypes"| node_c17
node_evaluator -->|"evaluates model"| node_network
node_evaluator -->|"reads settings"| node_config
node_exporter -->|"exports model"| node_network
node_exporter -->|"writes quantized model"| node_onnxmodel
node_network -->|"registers hooks"| node_gradcam

click node_app "https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/app.py"
click node_inference "https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/infer.py"
click node_bonefract "https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/dataset.py"
click node_c17 "https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/dataset.py"
click node_transforms "https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/dataset.py"
click node_trainer "https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/train.py"
click node_network "https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/model.py"
click node_gradcam "https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/model.py"
click node_evaluator "https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/evaluate.py"
click node_exporter "https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/export_onnx.py"
click node_onnxmodel "https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/models/bone_fracture_detector_int8.onnx"
click node_config "https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/configs/config.yaml"

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_app,node_inference,node_onnxmodel,node_cli toneBlue
class node_bonefract,node_c17,node_transforms,node_trainer,node_network,node_gradcam,node_config toneAmber
class node_evaluator,node_exporter toneMint
class node_patient toneIndigo
```

## Components

### Local diagnosis

- [Gradio application](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/app.py): `app.py`
- [Inference and overlays](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/infer.py): `src/infer.py`
- [INT8 ONNX model](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/models/bone_fracture_detector_int8.onnx): `models/bone_fracture_detector_int8.onnx`

### Model development

- [BoneFract loader](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/dataset.py): `src/dataset.py`
- [C17 subtype loader](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/dataset.py): `src/dataset.py`
- [Radiograph transforms](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/dataset.py): `src/dataset.py`
- [Training loop](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/train.py): `src/train.py`
- [Fracture classifier](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/model.py): `src/model.py`
- [Grad-CAM hooks](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/model.py): `src/model.py`
- [Pipeline configuration](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/configs/config.yaml): `configs/config.yaml`

### Validation and export

- [Test evaluation](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/evaluate.py): `src/evaluate.py`
- [ONNX export and quantization](https://github.com/deeptanu100/bone-fracture-crack-detection/blob/main/src/export_onnx.py): `src/export_onnx.py`

### Other components

- **Image submitter**
- **CLI inference user**

## Connections

- Image submitter → Gradio application: submits image (evidence: `README.md`)
- Gradio application → Inference and overlays: calls prediction (evidence: `app.py`)
- CLI inference user → Inference and overlays: runs inference (evidence: `README.md`)
- Inference and overlays → INT8 ONNX model: loads model (evidence: `src/infer.py`)
- Training loop → BoneFract loader: loads training data (evidence: `src/train.py`)
- Training loop → Radiograph transforms: uses transforms (evidence: `src/train.py`)
- Training loop → Fracture classifier: trains model (evidence: `src/train.py`)
- Training loop → Pipeline configuration: reads settings (evidence: `src/train.py`)
- Test evaluation → BoneFract loader: loads test data (evidence: `src/evaluate.py`)
- Test evaluation → C17 subtype loader: benchmarks subtypes (evidence: `README.md`)
- Test evaluation → Fracture classifier: evaluates model (evidence: `src/evaluate.py`)
- Test evaluation → Pipeline configuration: reads settings (evidence: `src/evaluate.py`)
- ONNX export and quantization → Fracture classifier: exports model (evidence: `src/export_onnx.py`)
- ONNX export and quantization → INT8 ONNX model: writes quantized model (evidence: `src/export_onnx.py`)
- Fracture classifier → Grad-CAM hooks: registers hooks (evidence: `src/model.py`)
