I'm grateful that someone is highlighting Keras and its usability and friendliness more! I never understood why people turned away from TF so much, I always found it a fantastic framework to work with as a researcher and professionally training models
Thank you so much, I really appreciate that! I’ve got in on the calendar haha, Keras Examples has object detection labeled as “Object Detection with RetinaNet” - Determined AI has a DETR tutorial up on their TH-cam channel. I will be “remixing” that one for my channel as well hopefully within a month or two.
Can this code also be used for 3D image classification from MRI scans? If yes, what normalization method will be appropriate? Nevertheless a very good tutorial. Good job!
Hi, thank you for your video! How to deal with not consistent amount of input data? Assume we have CT scans per patient, but there is can be 512x512x50, 512x512x300, 512x512x800 etc. If you have some links to read about it will be great. Or maby "def resize_volume" (more presisly the ndimage.zoom() function in it) in this example do exactly what im struggle to understand? :)
hi, you can do padding or do resampling(resampling will change 3D images voixel spacing, not safe)for not consistent data, in your case, in z dimension.
Yes! Francois Chollet has tweeted out a new one nearly every week and they are getting really impressive! Supervised Contrastive Learning was brought to Keras Example code in under a month. The documentation for TensorFlow and Keras is also getting really solid. Very easy to write custom loss functions, layers, callbacks, etc.
@@connor-shorten That's great. Thank you for going into detail with them. Btw, I presented the EfficientNet paper some days ago in my class and your video helped me with some details of it. Keep it up, mate!
I'm grateful that someone is highlighting Keras and its usability and friendliness more! I never understood why people turned away from TF so much, I always found it a fantastic framework to work with as a researcher and professionally training models
Thank you so much, I agree! I'm going to try to make some videos that highlight how PyTorch and TF/Keras are really similar right now
This is gold. Right here. I'm also wondering if has any company tried to build this in large scale and implement it healthcare industry...
Hey thanks for making this amazing walkthrough, I just came to know about this!
Thank you so much, really enjoyed making this! Hope to see more from you on Keras Code Examples!!
This is absolutely amazing ! Are you planning to also review the object detection example ? Many thanks for your time and effort :)
Thank you so much, I really appreciate that! I’ve got in on the calendar haha, Keras Examples has object detection labeled as “Object Detection with RetinaNet” - Determined AI has a DETR tutorial up on their TH-cam channel. I will be “remixing” that one for my channel as well hopefully within a month or two.
@@connor-shorten Nice ! Can't wait !
Great tutorial 👌 Thanks for sharing your knowledge 👍
Thank you so much.
Can this code also be used for 3D image classification from MRI scans? If yes, what normalization method will be appropriate? Nevertheless a very good tutorial. Good job!
Hey did you get to know about the normalisation method used for a MRI scan ?
Have you received any answer to this question?
Hello, have you found any insights on it?
Hi, thank you for your video! How to deal with not consistent amount of input data? Assume we have CT scans per patient, but there is can be 512x512x50, 512x512x300, 512x512x800 etc. If you have some links to read about it will be great. Or maby "def resize_volume" (more presisly the ndimage.zoom() function in it) in this example do exactly what im struggle to understand? :)
hi, you can do padding or do resampling(resampling will change 3D images voixel spacing, not safe)for not consistent data, in your case, in z dimension.
I want to execute the code in my system so please give the code
Are these keras code examples new? Did they just add them?
Yes! Francois Chollet has tweeted out a new one nearly every week and they are getting really impressive! Supervised Contrastive Learning was brought to Keras Example code in under a month. The documentation for TensorFlow and Keras is also getting really solid. Very easy to write custom loss functions, layers, callbacks, etc.
@@connor-shorten That's great. Thank you for going into detail with them. Btw, I presented the EfficientNet paper some days ago in my class and your video helped me with some details of it. Keep it up, mate!
That’s awesome to hear, thank you!
Can you give the code clearly