MIA - Microscopic Image Analyzer

Fast labeling with MIA utilizing segment anything

MIA is a software for deep learning based image analysis. It covers image labeling, neural network training and inference. It can be used for image classification, object detection, semantic segmentation and tracking.

Installation

The easiest way to install MIA is via conda (see https://docs.conda.io/en/latest/miniconda.html for installation options).

After installation of conda, download the environment file.

Then, open an anaconda prompt and type:

• cd /path/to/mia_environment.yaml (change /path/to/ to the path of the folder with the environment file)
• conda env create -f mia_environment.yaml
• wait and follow instructions

to start the software

type in an anaconda prompt:

• conda activate mia_environment
• mianalyzer

How to get help?

A quickstart guide can be found here and the complete user manual here.

Please use image.sc with the mia-tag for general discussion, questions about how to use the software or feature requests. Bugs can be reported directly in the issues panel on github.

Reference

If you use this code for your research, please cite:

https://www.cell.com/cell-reports-methods/pdf/S2667-2375(23)00146-7.pdf

Körber, MIA is an open-source standalone deep learning application for microscopic image analysis, Cell Reports Methods (2023)

Requirements

In general, MIA should run on any system with Linux or windows. You can use the cpu only, but it is highly recommended to have a system with a cuda-compatible gpu (from NVIDIA) to accelerate neural network training.