Hello Dr. Sreeni. Thank you for the great tutorial. I've got a WSI image dataset of tumors where there are multiple stain types, besides H&E. How do I detect these stain types and preprocess the images to make them uniform for the purposes of training a deep learning model to classify the tumor types?
It is really fantastic video. I highly appreciate your effort I have only one question, If i have a whole slide image with some annotation for multi object for segmentation, can i extract tile and extract masks for this annotated tiles for multi instance segmentation ? If yes how can i do that?
Hello Sir, thank you for the amazing tutorial. Can you please share the code for separating blank, partial and good images? It will be a great help to me.
Thanks again. I would appreciate it if you could provide me with your opinion on the "approximate number of patches that are enough for tumour detection in WSIs (e.g. 100k or millions or ...). I would like to know how many patches roughly should be extracted for training.
Nice tutorial. Following you from the beginning of the channel. Recently I am working with WSI. Used the same process to visualize the mask of a wsi which is in .tif format, but it displays black only. Need your guidance.
Amazing video!! I was wandering if I can save my trained model so I can call it one more time without rerunning it and how to do so. I'm actually trying to classify 3 images classes with a limited number of features like you have done in course number 158b and I'm getting an 0.59 in accuracy, I've tried data augmentation with ImageDataGenerator but the accuracy has become 0.58. What should I do to increase it. Thank you 🌹🌹
To save and load sklearn models you can use Pickle. import pickle #Save the trained model as pickle string to disk for future use filename = "sandstone_model" pickle.dump(model, open(filename, 'wb')) #To test the model on future datasets loaded_model = pickle.load(open(filename, 'rb')) result = loaded_model.predict(X) If you accuracy is not improving, play with the classifier parameters. If that doesn't help, you need more training data. You cannot augment your way to better accuracy.
Felicidades, es un buen ejemplo. 131 sentadillas son unos LIKESEX.Uno muchas y un buen ejercicio. Se deja ver que hay muy buenos resultados 😍👍 Saludos desde la Cd.. de world 🌹😉💖 los mortalesf abian apreciado tan hermosa mujer.k
You are my motivation in the medical imaging field, thanks Dr. Sreeni.
That was great. Thanks for your generosity in teaching.
My pleasure!
Will you have any separate videos on how to combine those small tiles and add them back to the original WSI?
Dr. Sreeni. thanks for interseted video but how we preprocess image all the fill data
Hello Dr. Sreeni. Thank you for the great tutorial.
I've got a WSI image dataset of tumors where there are multiple stain types, besides H&E. How do I detect these stain types and preprocess the images to make them uniform for the purposes of training a deep learning model to classify the tumor types?
It is really fantastic video. I highly appreciate your effort
I have only one question, If i have a whole slide image with some annotation for multi object for segmentation, can i extract tile and extract masks for this annotated tiles for multi instance segmentation ?
If yes how can i do that?
Thanks for this contribution!!!, do you know how to train models using tiles?.
Hello Sir, thank you for the amazing tutorial. Can you please share the code for separating blank, partial and good images? It will be a great help to me.
Thanks again.
I would appreciate it if you could provide me with your opinion on the "approximate number of patches that are enough for tumour detection in WSIs (e.g. 100k or millions or ...).
I would like to know how many patches roughly should be extracted for training.
Nice tutorial. Following you from the beginning of the channel. Recently I am working with WSI. Used the same process to visualize the mask of a wsi which is in .tif format, but it displays black only. Need your guidance.
Amazing video!!
I was wandering if I can save my trained model so I can call it one more time without rerunning it and how to do so.
I'm actually trying to classify 3 images classes with a limited number of features like you have done in course number 158b and I'm getting an 0.59 in accuracy, I've tried data augmentation with ImageDataGenerator but the accuracy has become 0.58.
What should I do to increase it.
Thank you 🌹🌹
To save and load sklearn models you can use Pickle.
import pickle
#Save the trained model as pickle string to disk for future use
filename = "sandstone_model"
pickle.dump(model, open(filename, 'wb'))
#To test the model on future datasets
loaded_model = pickle.load(open(filename, 'rb'))
result = loaded_model.predict(X)
If you accuracy is not improving, play with the classifier parameters. If that doesn't help, you need more training data. You cannot augment your way to better accuracy.
Felicidades, es un buen ejemplo. 131 sentadillas son unos LIKESEX.Uno muchas y un buen ejercicio. Se deja ver que hay muy buenos resultados 😍👍 Saludos desde la Cd.. de world 🌹😉💖 los mortalesf abian apreciado tan hermosa mujer.k
Thumb up.
Thanks
Hello.. very informative. CAn I connect with you over email plz to discuss my doubts about classification of H&E WSI? Kindly share email ID
Sorry, I get 100+ emails every day asking for help. I wish I had that kind of bandwidth to help everyone.