Hey maam, I am a final year medical student from India. This was a really insightful video. I myself am currently interested in the deep learning in radiology. I currently have access to more than 1 million radiology images. I want to use transfer learning method for my research, but this is too complex for me. That is why I am reaching out to you to let you know if you'd be interested in collaborating on my research project.
ma'am since the dataset contains only 2 classes i.e either tumor or nor then why are you using Categorical cross_entropy with 2 neurons why cant we use binary cross_entropy with 1 neuron?
thank you MOM . But i have one question. in this line of code vgg = VGG16(input_shape=(img_rows,img_cols,3), weights='imagenet', include_top=False), we are exercising on medical datset and they are not coloed image,we use the numbe rof channel=3 why? it is not clear for me. can you help me please.
Mam, img = cv2.resize(img,(224,224)) for these line I am getting error of (-215: Assertion failed) !ssize.empty() in function 'resize'. What to do?? Plz reply fast mam its my project of final year and tomorrow I have review for it.....plz help me.
can i show new imagen to this model so the model can tell me if a person have a tumor or not? since the dataset just have two seption ppl that have a tumor and people that doesnt have a tumor, it will be cool if you add imagen withour any context to see if the model can guest if a person have a tumor or no.
Hello. Hope you are doing fine. I am an undergrad student . I am working on a project. Can you guide me in a query. I have CT scan dataset in the form of DICOM meaning for a single CT scan I have 100 layers. Should I feed the whole data into my deep learning model or should I feed the single malicious layer into my model. Thankyou
Why did you add layers to the model? Is it required for transfer learning to work or is it just an attempt at improving accuracy? Also, do you know if it is possible to display a class activation map, and how to do it?
Those models are built on ImageNet dataset. If you want to create model for your dataset then we need to add some top layers so that it learns the specific features of ur dataset. Otherwise accuracy may be very low.
Hey maam, I am a final year medical student from India. This was a really insightful video. I myself am currently interested in the deep learning in radiology. I currently have access to more than 1 million radiology images. I want to use transfer learning method for my research, but this is too complex for me. That is why I am reaching out to you to let you know if you'd be interested in collaborating on my research project.
Ya sure. This is my mail id: hemalatha2107@gmail.com
I would like to collaborate too
Very informative. Please make such videos for image segmentation as well. Thanks
Superb session ..i will be in touch with you for my research work . Your help is appreciated mam
Sure 👍
very less time to cover more helpfull info , thanks u
thank you so much mam.. super clear cut explanation mam..kindly upload image to text genration also from generative ai mam kindly
Thank You So Much! well explained.!. can you make more videos about deep learning
Good job mam keep vloging God bless 🙏
ma'am since the dataset contains only 2 classes i.e either tumor or nor then why are you using Categorical cross_entropy with 2 neurons why cant we use binary cross_entropy with 1 neuron?
Thank you mem...thanks a lot...❤️
Most welcome 😊
thank you MOM . But i have one question. in this line of code vgg = VGG16(input_shape=(img_rows,img_cols,3), weights='imagenet', include_top=False), we are exercising on medical datset and they are not coloed image,we use the numbe rof channel=3 why? it is not clear for me. can you help me please.
is she your mom? that's so cool..
Amazing tutorial
Thank You So Much! well explained.!
how to Identify whether an image has a tumor or not by uploading an image as an input?
Mam, img = cv2.resize(img,(224,224)) for these line I am getting error of (-215: Assertion failed) !ssize.empty() in function 'resize'. What to do?? Plz reply fast mam its my project of final year and tomorrow I have review for it.....plz help me.
Did the problem resolve? Tell me i have practical today
nice video, one note 100 - 33 = 67, not 77, at 6:51
12.48 time , why softmax activation for two class problem, can we use sigmoid there?
Pleaselease visit my channel and patiently listen about vgg19 ur doubts all doubts will be cleared. Sigmoid cannot be it's not adding upto..
dear brother sigmoid activation function can be used.
Softmax is good for classification.. it enhances the true positives.
Softmax is like probabilities adding up to 1
Since it is 2 class classification problem so you can use sigmoid . Suppose it is 3 class classification then you've to use softmax .
can i show new imagen to this model so the model can tell me if a person have a tumor or not? since the dataset just have two seption ppl that have a tumor and people that doesnt have a tumor, it will be cool if you add imagen withour any context to see if the model can guest if a person have a tumor or no.
Thank you Madam. Which transfer learning model will you suggest for biological sequence classification ?
ProtBERT
Hello. Hope you are doing fine. I am an undergrad student . I am working on a project. Can you guide me in a query. I have CT scan dataset in the form of DICOM meaning for a single CT scan I have 100 layers. Should I feed the whole data into my deep learning model or should I feed the single malicious layer into my model. Thankyou
thank you for this
Very good 👍👍
Pleaselease visit my channel and patiently listen about vgg19 ur doubts all doubts will be cleared.
why is the color channel of the model set to 3 if the images are in grayscale? Shouldn't it be 1?
Where can I find the code which you have written ma'am
Super presentation. One question, Can we add xgboost classifier into this model?
Yes u can
Why did you add layers to the model? Is it required for transfer learning to work or is it just an attempt at improving accuracy?
Also, do you know if it is possible to display a class activation map, and how to do it?
Those models are built on ImageNet dataset. If you want to create model for your dataset then we need to add some top layers so that it learns the specific features of ur dataset. Otherwise accuracy may be very low.
dear madam why you didn't detected tumor in images,?? i'm interested for this model so please answer me and thnks
Well explained
Pleaselease visit my channel and patiently listen about vgg19 ur doubts all doubts will be cleared.
what accuracy does this model get
Please, how can i pass my dataset without training in Vgg16(FC7-Layer- Feature extraction) through Jupyter Notebook?
if you are going to use only for feature extraction then you can set layer.trainable=False
@@investime247 Thank you so much
Can We use the same model for accident detection images (yes/no)
Ya sure
can we have the google collab fille please ?
Might be model is overffited!