🫶🏻 Notion of the repo: https://blossom-rayon-fd7.notion.site/Tesis-leucemia-8c9f204848e64724a6d60bef59e76a95?pvs=74 (is in spanish)
Welcome to the Leukemia Classifier project! This initiative employs machine learning, computer vision, and various tools to classify Acute Lymphoblastic Leukemia (ALL) images into distinct categories.
This project utilizes Flask, TensorFlow, and scikit-image to offer a user-friendly interface for leukemia image classification. It integrates two machine learning models:
One for distinguishing between 'hem' and 'all' leukemia types. Another for classifying the leukemia stages: 'Benign', 'Early', 'Pre', and 'Pro'.
Image Classification: Upload leukemia images and classify them into specific types and stages. Model Integration: Combines machine learning models to provide accurate classification results. User Authentication: Secure login for authorized access to diagnostic features.
To set up and run this project locally, follow these steps:
Clone the repository:
git clone https://github.com/nicolasvargaszz/acute-lymphoblastic-leukemia-classifier.git cd leukemia-classifier
pip install -r requirements.txt
Run the Flask app:
python app.py
Access the app via http://localhost:5000 in your web browser.
Home/Landing Page: Access the landing page showcasing project details and functionalities. Login: Authenticate as a doctor by entering valid credentials. Upload & Classify: Upload leukemia images for classification into specific types or stages. Results & Database: View classification results and stored data in the app's database.
Flask: Web framework for developing the application. TensorFlow: Framework for building and deploying machine learning models. scikit-image: Library for image processing and computer vision tasks. SQLite: Database for storing classification results and other data. NOTE:
ALL: Acute Lymphoblastic Leukemia (Spanish: Leucemia linfoblastica aguda) HEM: Cells without leukemia
Here are some examples of the classification results:
.jpeg)
For any questions or further information, feel free to reach out:
Twitter: @nicoelingeniero