Standalone versions of BONe DLFit, BONe DLPred, and TorchCUDA IoU (Amira-Avizo is not needed)

Citation: Lee et al. in prep.

To fit a 2D, 2.5D, and 3D models, use: BONe_DLFit_venv.py To perform AI segmentation, use: BONe_DLPred_venv.py To calculate IoU Score, use: TorchCUDA_IoU_venv.py

Installation and operation: Windows (tested on 10/11 Pro)

1. Install Python 3.12 (tested 12.0, 12.3, 12.10) for full PyTorch support (as of 10/14/2025).

2. Copy BONe_apps folder to computer.

3. Double-click to run: BONe_DLFit_venv.py (train a model) or BONe_DLPred_venv.py (apply AI model to segment a micro-CT scan) or BONe_IoU_venv.py (evaluate predicted segmentation)

   Note: step 3 works as intended on your computer if py files are associated with Python 3.12. If not, right-click the desired py file and Open with Python 3.12. Alternatively, open a Terminal or Powershell window from the BONe_apps folder. [if Python 3.12 is the only version on the computer] Type: py BONe_DLFit_venv.py [if there are different versions of Python on the computer] Type: python3.12 BONe_DLFit_venv.py

4. Each app will check for Python 3.12, pip, and PyTorch (2.8 as of 10/14/2025) before running. Upon initial installation, user will be asked to choose a CUDA version of PyTorch (12.9 for support up to 50-series/Blackwell Nvidia GPUs; 12.6 for support up to 40-series/Ada Lovelace/Hopper Nvidia GPUs; or CPU for systems without an Nvidia GPU).

5. Close app window to exit.

Installation and operation: Ubuntu Linux (tested on 22.04 LTS)

1. • Install Python 3.12 (tested 12.0, 12.3, 12.10) for full PyTorch support (as of 7/12/2025)

2. Copy BONe_apps folder to computer (drive must be in ext4 format).

3. Open Terminal at location of BONe_apps folder, and type: python3.12 BONe_DLFit_venv.py (training a model) or python3.12 BONe_DLPred_venv.py (apply AI model to segment a micro-CT scan) or python3.12 BONe_IoU_venv.py (evaluate predicted segmentation)

4. Each app will check for Python 3.12, pip, and Pytorch (2.8 as of 10/14/2025) before running. Upon initial installation, user will be asked to choose a CUDA version of PyTorch (12.9 for support up to 50-series/Blackwell Nvidia GPUs; 12.6 for support up to 40-series/Ada Lovelace/Hopper Nvidia GPUs; or CPU for systems without an Nvidia GPU).

5. Close app window to exit.

• (Recommended method of installing Python using Ubuntu 22.04 LTS) sudo add-apt-repository ppa:deadsnakes/ppa sudo apt update sudo apt install python3.12{,-venv,-tk, -dev, -distutils}