doors-and-windows

More details into the report is present under report.pdf

A lightweight web application that detects doors and windows in architectural floor plans.

This application has been built using Streamlit for the front-end and supports three YOLO variants for the prediction, namely:

• YOLOv8n
• YOLOv8s
• YOLOV8m

Features:

• Upload blueprint images in the form of either '.jpg', '.jpeg', or '.png'.
• Select from different YOLOv8 variants.
• Visual detection output along with bounding boxes.
• Output includes label, confidence scores, and bounding box coordinates.

Setup Instructions

1. Clone the repository

git clone https://github.com/akshathmangudi/doors-and-windows.git
cd doors-and-windows/

Create and activate a virtual environment

For virtualvenv

python -m venv <env_name> 
source <env_name>/bin/activate # On Windows use 'venv\Scripts\activate

For conda

conda create -n <env_name> python=3.10
conda activate <env_name>

Install Dependencies

pip install -r requirements.txt

Run the Streamlit App

streamlit run app.py

This app will launch in your browser at http://localhost:7860

To run the API Locally (FastAPI Example)

If you're running the backend as an API:

uvicorn src.main:app --host 0.0.0.0 --port 8000 --reload

And view the API in http://0.0.0.0:8000/docs

Linting and Formatting

To ensure the code has been maintained properly and professionally, we use pylint, black and isort for linting, formatting and import sorting respectively.

Example API Usage

To test the public API:

Go to the link that's present in the "About" section or click here.

Endpoint

POST /detect
Content-Type: multipart/form-data

Request Fields

Field	Type	Description
image	file	Floorplan image (JPG/PNG)
model_name	string	Model variant (e.g. yolov8n)

Example Response

{
  "filename": "image.jpg",
  "detections": [
    {
      "label": "door",
      "confidence": 0.848,
      "bbox": [852.94, 467.47, 975.9, 613.08]
    },
    {
      "label": "door",
      "confidence": 0.842,
      "bbox": [378.89, 268.19, 504.86, 414.77]
    },
  ],
  "total_detections": 2,
  "confidence_threshold": 0.5,
  "image_shape": [1021, 1143, 3]
}

Project Structure

├── app.py
├── config.yaml
├── contract.yaml
├── Dockerfile
├── __init__.py
├── LICENSE
├── model_comparison.txt
├── models
│   ├── onnx
│   │   ├── yolov8m.onnx
│   │   ├── yolov8n.onnx
│   │   └── yolov8s.onnx
│   └── pt
│       ├── yolov8m.pt
│       ├── yolov8n.pt
│       └── yolov8s.pt
├── README.md
├── requirements.txt
└── src
    ├── config.py
    ├── detection.py
    ├── __init__.py
    ├── main.py
    └── utils.py

License

See the LICENSE file or more info.