Algorithm to analyze perfusion (flow) of capiliary blood vessels using machine learning
- Make sure you have the following addons:
- Matlab runtime 2020a
- Computer vision add-on
- Image processing add-on
- Make sure you have python 3.5-3.7 installed
- To check python version:
- MAC: Open spotlight search, type terminal, open terminal window
- WINDOWS: open start menue, search for cmd, and open cmd terminal
- in terminal write: python3 --version
- Proceed if in the terminal a version number like 3.7.X shows up
- Otherwise install python from here. https://www.python.org/downloads/release/python-377/
- Download or clone repository
- Run pip install -r requirements.txt in the project directory
- Now you can run the VesselDetector.exe
To use APPLICATION:
- Press browse button, then select target Experiment video file (.avi)
- When the video appears, you may press play button to see it play, then pause
- After select settings for analysis:
- Number of crosssections: 3 to 10 cross-sections
- Flowing threshold: increase to be detect fewer borderline vessels
- Min vessel length: set minmum length in pixels of vessel to be accepted
- (Probably should not change this too much)
- then press analyze results
- Be patient analysis can take between 2-4 minutes depending on computer speed
- See results in new window :()
- Results are modifiable:
- To remove a point detected by the computer, just click on it
- To add a point, just click on the image where you would like to add it
- When ready press save results to save results
- Enter prefix name of files to save:
- Program will save 3 files with entered prefix
- prefix_initial_points_found.txt
- prefix_final_points_found.txt
- prefix_SAD_results.bmp
- for intermediaray results, look through working directory
- Results are modifiable: