AIM

A lightweight wrapper around ITK ScancoIO to read, convert, and align .aim files from Scanco microCT scanners.

Installation

This package requires Python 3.8 due to ITK-ScancoIO compatibility constraints.

Installation

git clone https://github.com/wallematthias/ScancoAIMreader.git
cd ScancoAIMreader
conda env create -f environment.yml --name aimenv
conda activate aimenv
pip install -e .

Overview

This repository provides Python utilities for:

• Reading .aim files into AIMFile objects
• Converting Hounsfield units to density (mg/cm^3)
• Writing AIM files back to .mha
• Aligning multiple AIM scans to a shared coordinate space

It is designed to support legacy workflows using Scanco AIM files, such as preclinical bone imaging pipelines.

Usage

Load an AIM file

from aim import load_aim

aim = load_aim("sample.AIM")
print(aim.data.shape)
print(aim.voxelsize)

Write an AIM file

from aim import write_aim

write_aim(aim, "output.AIM")

Align multiple AIM files to a common space

from aim import pad_to_common_coordinate_system

padded_arrays, common_origin = pad_to_common_coordinate_system([aim1, aim2])

Class: AIMFile

An in-memory representation of an AIM scan:

• data: A 3D pint.Quantity array (e.g., in mg/cm^3)
• voxelsize: A pint.Quantity with voxel size in mm
• position: Origin as voxel index (in scanner coordinates)
• processing_log: Metadata from the AIM header

Functions

• load_aim(path): Reads and converts .AIM or binary mask files.
• write_aim(aim_file, path): Writes an AIMFile object to disk.
• pad_to_common_coordinate_system(aim_files, padding_values=None): Pads multiple AIMFile objects to a shared coordinate system.
• get_aim_calibration_constants_from_processing_log(path): Extracts calibration constants from raw AIM headers.
• convert_hounsfield_to_mgccm(...): Converts Hounsfield units to bone mineral density using scanner-specific calibration.

Notes

• This repository assumes familiarity with Scanco microCT data and is not intended as a general-purpose AIM file reader.
• The load_aim() function handles both image data and masks.
• You can easily integrate it into preprocessing workflows for morphometry or biomechanical analysis.