ai-center 2026.5.6

Automated Sample Alignment for MX using a YOLO Model

aicenter

A python based Soft IOC Server for Sample Alignment using a YOLO model.

Usage

python -m venv my-venv
source my-venv/bin/activate
pip install ai-center

1. Create a directory for the IOC instance.
2. Copy the start.sh file from the deploy directory into this directory.
3. Copy the ai-centering.service unit file from the deploy into your /etc/systemd/system directory.
4. Edit the files from (2) and (3) above to reflect your environment and to set all the required instance parameters
5. Enable the unit file using your system commands. For example, systemctl enable ai-centering.
6. Start the init file using your system commands. For example systemctl start ai-centering.

You can manage the instance daemon through procServ, by telneting to the configured port.

OpenCV

For best performance, a version of python-opencv compiled with support for CUDA and cuDNN along with a compatible GPU should be used.

Testing

The test/inference.py file can be used to test the inference / model performance without running a full IOC application. To obtain a copy of the trained model weights, contact the authors.

Create an environment with requirements-test.txt and install aicenter without dependencies:

pip install -r requirements-test.txt
pip install . --no-deps