MIPAR

MIPAR is an image analysis software package that enables users of all experience levels to create advanced segmentation algorithms for extracting quantitative information from images. Developed in the United States, originally by researchers associated with Ohio State University, MIPAR, short for Materials Image Processing and Automated Reconstruction, was first built for materials science and has since expanded to life sciences, geology and other fields. The problem it addresses is that measuring features in microscope and other scientific images, such as grains, particles, pores, phases or cells, is often done manually or with generic tools that struggle with complex, noisy images. Writing custom code requires programming skills many scientists and engineers lack. MIPAR provides a visual, recipe-based environment in which users combine processing steps into reusable workflows, supported by deep learning tools that can learn to segment difficult features from a small number of annotated examples.

The software includes a large library of image processing filters, thresholding, morphological operations and measurement functions, all arranged into recipes that can be saved, shared and run in batch across thousands of images. The Deep Learning Trainer enables users to train neural networks by annotating examples, achieving robust segmentation of complex structures without coding. Measurement outputs include size, shape, area fraction, orientation and spatial distribution, with statistical summaries and exportable reports. Specialised modules support standardised analyses such as grain size measurement according to ASTM standards, porosity analysis and 3D reconstruction from image stacks. MIPAR runs on Windows and macOS and can also be integrated into automated workflows. It is used by materials scientists, metallurgists, quality control laboratories, additive manufacturing groups, universities and research institutes in industries including aerospace, automotive, energy and biotechnology. Pricing starts from around USD 4,950 with a free trial, and a free version is also available.