myovision-app

Description

This is a sub-project of the main project myovision. It provides a Backend for the Application developed for the convenient use and further development of the myovision project. It is a RESTful API developed with the FastAPI framework and has a designated Frontend for the user interaction myovision-app-front

With a slight modification the Backend can be generalized and used for any computer vision project concerning instance segmentation.

Visualizations

• Annotation Tool for collecting labeled data Video

• Inference Tool for observing model's predictions and relevant metrics Video

Setup

To setup the Application backend locally follow the steps:

1. Get Model Checkpoint

• Get the model checkpoint from link
• Place it in the ./checkpoints/ directory

2. Install Dependencies

• Install python dependencies using: pip install -r requirements.txt
• Install redis: follow the simple instructions based on your OS link

3. Set Environment variables

• Copy the example .env file: cp .env-example .env
• Adjust the default variables in .env file if desired

Starting Application

1. Start Redis: follow the simple instructions based on your OS link

2. Start API:

   uvicorn backend.main:app

3. Start Frontend: Follow the Instructions

Finally, open the browser go to the url the frontend is running.

Using Container Images

Both Backend and Frontend contian Dockerfile to build docker images using: docker build (docs).

it can then be conveniently managed with docker-compose: docker-compose up (docs).

Contact Information

{
    name = "Giorgi Nozadze",
    email = "giorginozadze23@yahoo.com"
}