Document Classifier (Estácio/UNIASSELVI) and Answer Sheet Identifier ✅

This application classifies uploaded documents (PDFs, DOC/DOCX, or images) as:

• Estácio Exam
• Uniasselvi Exam
• Other

It also detects checkmark boxes using a YOLO model, displaying them with green bounding boxes and showing statistics like count and average confidence.

Features

• File Conversion: DOC/DOCX to PDF, PDF to images.
• Document Classification: Uses a LeViT model (ONNX format) to identify the type of document.
• Checkmark Detection: YOLO model identifies checkmarks with confidence thresholding and non-maximum suppression.
• Streamlit Interface: Clean UI with sidebar model selection, image previews, and detection stats.

Installation

System Requirements

• Python 3.10+
• Poppler for pdf2image
• LibreOffice (for DOC/DOCX conversion)

Install dependencies (after cloning the repository):

sudo apt update
sudo apt install poppler-utils libreoffice -y
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Running the App

streamlit run app.py --server.fileWatcherType none

Access the app at http://localhost:8501.

Running on AWS EC2

1. Launch an EC2 instance with Ubuntu.
2. Install dependencies as described above.
3. Clone this repository and navigate into it.
4. Start the Streamlit app with:

streamlit run app.py --server.headless true --server.port 8501 --server.enableCORS false

5. Open your browser to http://<your-ec2-public-ip>:8501.

Folder Structure

.
├── app.py
├── requirements.txt
└── models/
    ├── levit-384-exams-classification/
    │   ├── onnx/levit384_estacio-multiclass.onnx
    │   ├── preprocessor_config.json
    │   └── metrics/metrics.png
    └── yolo_detection_checkmark-final/
        └── best.pt

License

Apache License

---

Developed by Lucas Meireles