potpov/ian_annotation_tool

Tool for annotating the alveolar canal from 3D images. Source code for the paper: "A cone beam computed tomography annotation tool for automatic detection of the inferior alveolar nerve canal. VISAPP 2021"

IAN Annotation Tool

This tool is designed to help in the annotation of the inferior alveolar nerve (IAN) canal in cone beam TAC. It is developed using:

• PyQt5 (GUI)
• mayavi (3D visualization)
• opencv, scikit (Image processing)

Run with Python interpreter

Create virtual environment and install requirements via pip. It's important to use pip instead of conda packages in order to package the application into an executable.

python -m virtualenv venv
source venv/bin/activate
pip install -r requirements.txt

On Linux

It is required to install PyQt5 on the system.

sudo apt-get install python3-pyqt5

Build executable

What follows is the configuration used to freeze the application into an executable.

	Version
OS	Windows 10 2004 build 19041.508
Python	3.7.3
cx_Freeze	6.1

1. Setup virtual environment from requirements.txt:

   python -m virtualenv venv
   source venv/bin/activate
   pip install -r requirements.txt

2. Add DLLs to virtual environment. DLLs are available in Python installation, so navigate to that folder, copy DLLs directory and paste it into venv.
3. Launch build_exe.py:

   python build_exe.py build

   Instead, to create an installer for Windows, run the following command:

   python build_exe.py bdist_msi

4. Check into build directory.

Precalculate titlted planes and images

To precalc the tilted planes and images of side volume given a set of DICOMs pre-annotated by technicians, you need to use tsv_precalc.py.

usage: tsv_precalc.py [-h] -d DIR [-f] [-c] [-w WORKERS]

optional arguments:
  -h, --help  show this help message and exit
  -d DIR      Directory to explore to find DICOMDIR
  -f          Force re-computation even if side volume is already available
  -c          Clean directory from saves and other data
  -w WORKERS  Amount of workers for concurrent side volume computation

Export gt_volume.npy, masks.npy and imgs.npy

To export gt_volume.npy, masks.npy and imgs.npy given a set of DICOMs, you need to use annotation_export.py.

usage: annotation_export.py [-h] -d DIR [-f] [-w WORKERS]

optional arguments:
  -h, --help  show this help message and exit
  -d DIR      Directory to explore to find DICOMDIR
  -f          Force re-computation even if gt_volume.npy already exists
  -w WORKERS  Amount of workers for concurrent extraction