Thanks for your informative solution! How did you get the QuPath0.4.0 OpenCV JavaFX Application Thread command window opened?
@pataporon
Ай бұрын
There should be an executable in the QuPath folder to "Launch with console". If you can't find the executable, then it might only be generated when building from source, as I had.
@darktempalor
Ай бұрын
Hello, thank you this has been super helpful! I'm currently trying to do something just like this with pictures I have taken on a scanner of 4 plates at a time, I'm trying to figure out how to get Qupath to calculate each petri dish Seperately using this method, would you be able to point me in the right direction?
@pataporon
Ай бұрын
Sure! Perhaps the easiest approach would be to split the plates before loading them into QuPath. Lots of image editing programs like MS Paint can be used to crop out regions. Or if you want something for non-rectangular cropping, then GIMP could be an alternative. Once the plates are separated, each to their own image, you can proceed from there. Alternatively, if you wanted to keep the images as-is, the next best option would be to create 4 separate annotation regions per image. You'd follow the steps in the video up until the pixel classifier is created. Then, you can apply the pixel classifier to all annotations (4) in the image. Applying it using the "measure" button, will give you the percentage area classified for each class used in training the classifier. Then, you can export annotation measurements, where each annotation corresponds to a single petri dish, with the area as the columns. It is the more complicated of the two approaches, which is why I'd recommend splitting the plates before loading into QuPath.
@darktempalor
13 сағат бұрын
@@pataporon Thank you! This helped alot, currently trying to train the pixel classifier to detect stained Hyphae as well! Qupath rocks!
Пікірлер: 5