Frequently Asked Questions

This page lists common issues encountered during the installation and use of MIRA, along with their solutions.

Installation & Environment

🛠️ Qt Platform Plugin “xcb” Error (OpenCV Conflict)

Issue: On Linux, MIRA fails to start with the error: Could not load the Qt platform plugin "xcb" in ".../cv2/qt/plugins". This is caused by a conflict between the Qt libraries bundled with opencv-python and those used by MIRA’s interface.

Solution: You must force the use of the “headless” version of OpenCV, which does not include its own Qt binaries:

pip uninstall opencv-python
pip install opencv-python-headless --force-reinstall
🚀 NVIDIA GPU not detected

Issue: MIRA runs on CPU even though a GPU is available.

Solution: Ensure PyTorch is installed with CUDA support. Check your version with nvidia-smi and reinstall accordingly:

# Example for CUDA 11.8
pip install torch torchvision --index-url https://download.pytorch.org/whl/cu118

Performance & Usage

📉 Out of Memory (OOM) Error

Issue: Crashes during high-resolution image processing.

Solution: * Reduce Image Size: Lower the input resolution in the settings. * Hardware: Ensure your GPU has at least 4GB of VRAM.

🔍 Low Detection Accuracy

Issue: Model fails to detect objects or has too many false positives.

Solution: * Thresholds: Adjust the Confidence and IoU parameters in the UI. * Quality: Ensure microscopy images are well-focused and have consistent lighting.

Still having issues? If your problem is not listed here, please open an issue on the Issue Tracker.