Skip to content

API for multi-class classification of ultrasound images. The JSON response allows to print the probability of each class

License

Notifications You must be signed in to change notification settings

M-ballabio1/API-Ultrasound-Classificator

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

38 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

API-Ultrasound-Classificator 🩺

Details

This project is an API classifier built using a deep learning architecture. It allows you to classify various types of data with high accuracy. You can interact with the application through a user-friendly Streamlit app. Check out the live demo at App Classifier. The code and detailed documentation can be found in this repository.

VERSION V1▶️

This is the workflow and pipeline utilized in version 1 of this API. The process involves multiple steps, including data collection, preprocessing, feature engineering, model training, validation, and deployment. Each stage is carefully orchestrated to ensure the seamless functioning of the API, delivering efficient and accurate results to users.

mlops_pipeline_versione1

This is the screenshot of web application used a default client of this api.

web_app_st

[Live Demo App: App Classifier]

Repository

The code and detailed documentation can be found in this GitHub repository.

Features

  • FastAPI with deep learning model of classification
  • Hosting API and monitoring access using Render.com
  • Deep learning architecture for accurate classification
  • User-friendly interface powered by Streamlit

Usage

  1. Install the required dependencies.
  2. Clone the repository.
  3. Run the Streamlit app.
  4. Upload your data for classification.
  5. View the classification results.

For more detailed instructions, refer to the documentation in the repository.

Other usage on: [Live Demo App: App Classifier

Image 1: Interaction using Streamlit UI.

Interaction with the API:

POST: https://api-ultrasound-image-classificator.onrender.com/classification

  • Input: images (png or jpg)
  • Output: json string (str with odds of classification in each classes)

Render-API Image 2: Render platform to monitor Https requests sent.

VERSION 2▶️

This is the workflow and pipeline utilized in version 2 of this API. The process involves multiple steps, including data collection, preprocessing, feature engineering, model training, validation, and deployment. Each stage is carefully orchestrated to ensure the seamless functioning of the API, delivering efficient and accurate results to users. There are different approach based on automatization of process using Github actions, push in DockerHub and GCS and deploy on Cloud Run (GCP).

mlops_pipeline_versione2

This is the screenshot of web application used a default client of this api.

web_app_st

[Live Demo App: App Classifier]

Repository

The code and detailed documentation can be found in this GitHub repository.

Features

  • Automatically detection of changes in repository and push in Docker Hub, Google Cloud Registry and deploy in Cloud Run.
  • FastAPI with deep learning model of classification
  • Hosting API and monitoring access using Cloud Run (GCP)
  • Deep learning architecture for accurate classification
  • User-friendly interface powered by Streamlit

Usage

  1. Install the required dependencies.
  2. Clone the repository.
  3. Run the Streamlit app.
  4. Upload your data for classification.
  5. View the classification results.

For more detailed instructions, refer to the documentation in the repository.

Other usage on: [Live Demo App: App Classifier

Image 1: Interaction using Streamlit UI.

Interaction with the API:

POST: https://api-ultrasound-classificator-cloud-run-pa6vji5wfa-ew.a.run.app/classification

  • Input: images (png or jpg)
  • Output: json string (str with odds of classification in each classes)

cloud_run_api

Image 2: Cloud Run Api

Feedback and Contributions

Feel free to provide any feedback or suggestions by creating an issue in the repository. Contributions are also welcome!

License

About

API for multi-class classification of ultrasound images. The JSON response allows to print the probability of each class

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages