Computer Vision News - March 2021

19 ImageJ Another option for use of ImageJ in combination with Python language is Jython scripting, a Java implementation of Python. Unfortunately, this is only currently supportingPython2and it can’tuseexternal Pythonmodules suchasNumPy.However, the JavaCPP API, including almost all functions of OpenCV, can be employed. In this way, Python scripts can be used as customised plugins for your ImageJ or bundled in perfectly automated packages Conclusion ImageJ is a wide-ranging software offering hundreds of image processing algorithms and analysis tools of medical images, and not only! It is well-known for its easy-to-use interface, the specific plugins and macros to automate processes by recording users’ interactions. In combination with other powerful libraries such as OpenCV, it can also effectively perform automatic and manual tasks- such as finding ROI in images and then manually adjusting them. Through the use of external libraries, plugins, and wrapper functions for integration between ImageJ and other software packages, it can be included in complex computer vision and machine learning pipelines, making it a uniquely versatile program. # Convert image to suitable format to apply ImageJ filter result = ij.py.new_numpy_image(cell_colony) imgIterable = ij.op ().transform().flatIterableView(ij.py.to_java(cell_colony)) resIterable = ij.op ().transform().flatIterableView(ij.py.to_java(result)) # Add poisson noise to image ij.op ().filter().addPoissonNoise(resIterable, imgIterable) # Show result ij.py.show(result ,cmap='gray') # Clip the result (show only noise) result = np.clip(result, 0 , 45 ) ij.py.show(result.astype(np.uint8),cmap='gray')