Detect COVID19 from chest CT-scan. Deep Learning algorithm to detect corona from chest CT images.

[@andreiramboiu5184]

Hello, the model link and data set link not working

[@karanpatel3358]

New Dataset link: drive.google.com/file/d/1hUjxS2LfOmugvqIGhcKd5G2J1jmRjByi/view?usp=sharing New Model Link: drive.google.com/file/d/1DdKCPLf3JM1rzPZko3QwjTzisdU9hCq6/view?usp=sharing

[@viper-ms1us]

Nice

[@mmm-eo5cs]

Thankyou sooo much sir .. you don't know how helpful this video is..... Thanks alot... Sir mera bs ek question hai..... Humein model ki overall accuracy kese pta chlegi???? Agr train krte time aapko mili ho to plz btaiye ..

[@karanpatel3358]

Thank you for your support keep sharing the video. There are multiple code from which you can determine the overall accuracy however you should also consider ROC curves and other graphs.

[@mmm-eo5cs]

Ok sir..... You have written 95percnt accuracy in thumbnail..... Aap plz btayenge aapne konsa parameters consider kiye hain...

[@karanpatel3358]

@@mmm-eo5cs it’s general KPI from tensorflow which gives you accuracy.

[@mmm-eo5cs]

@@karanpatel3358 okk thankyou soo much sir..... And thankyou for quick response... 🙏🏼🙏🏼

[@medicalengineering35]

Nice work !!!

[@karanpatel3358]

Medical Engineering thank you so Much 😊

[@arthpatel5688]

Wow!! great work

[@karanpatel3358]

Arth Patel thank you so much 😊

[@VijayaLakshmiRajanna-yl2fo]

In vgg classifier what are all the parameters it taken into considerations to say that the dataset image is +ve to virus?

[@karanpatel3358]

As far as the classification of pre-train data; physicians already classified into +ve and -ve categories.

[@abhijeetanand3443]

how to open .h5 vgg file to make changes in certain layer?

[@farahjabeen7707]

@Karan Patel I want yo learn about medical image processing using deep learning. I don't know about free resources where i can learn about medical images and how to use them in deep learning. Can you guide me on a little pathway. Thank you for your time in advance

[@decimalinfotech1060]

hello sir, could you help me to find the masks of every ct images or roi using python

[@davisjohn-d6h]

Hi Karan, don't you have to do model.fit in order to train the model? I don't see model.fit here unless I missed it somewhere. And also don't you have to shuffle the data at the time stamp 6:41 when you do np.concatenate, because the order of X_train is first all noncovid images and then comes all covid images. Don't you have to shuffle the X_train and X_test in order to randomize the order of noncovid and covid? Because if all the noncovid images come first then the model can just simply predict noncovid? Thanks

[@karanpatel3358]

First of all Davis thank you for comment, It is not like if you don't shuffle the dataset you can not predict another side. Shuffling data ensuring that each of your data gives independent impact on model. I hope you are clear with your doubt.

[@davisjohn-d6h]

@@karanpatel3358 Thanks for the reply. Also, don't you have to do model.fit to train the model? I didn't see a model.fit here, I just saw you predict

[@karanpatel3358]

@@davisjohn-d6h because model.fit command for to feed the data in dense layer. To shorten the code “I run code in my computer for 50 epochs and directly read that model” Thank you

[@KaranPatel-wc4wo]

@@karanpatel3358 thank you sir

[@susmitajkumar5441]

can you explain the vgg model that you have used here

[@bouchrarahmani4225]

good job , how can i emerge or fusion between CT scan and Xray image.?

[@karanpatel3358]

I think you can’t because these image capture from different modalities.

[@bouchrarahmani4225]

The problem in this. I have a final project about fusion between ct image and xray with deep learning

[@saisrujan6918]

Bhai which algorithm did you use

[@karanpatel3358]

Deep learning algorithm

[@saisrujan6918]

@@karanpatel3358 thank you for your reply

[@karanpatel3358]

Any time, Code is available on github.

[@shivtejful]

Can images with different pixel spacing inputed as arrays in training have effect on model generation/performance?

[@karanpatel3358]

Yes, pixel spacing does matter on model training as far as my experience says. But I manage my pixel spacing between Pixel Spacing = Row Spacing \ Column Spacing = 0.30\0.25. However i did not set this feature in this algorithm. Thank you .

[@shivtejful]

What are the ways pixel spacing can impact the model according to your experience?

[@karanpatel3358]

Let's Think majorly when the pixel spacing is ideal it’s get more proper Num of features. If you it is spacing is larger then it will get less features and it may spoil your performance.

[@davisjohn-d6h]

It's hard to hear you with that background music, other than that, good video

[@karanpatel3358]

Okay sorry for that, in next video I will resolve this ! Thank you for the feedback Davis John

[@譚聯偉]

i cannot use the tensorflow and matplotlib simutaneously.... how can i do ? thanks a lot >

[@karanpatel3358]

譚聯偉 try the code which I put in description.

[@debkumar5536]

Can you please share the vgg model code with me

[@karanpatel3358]

Deb Kumar I did not get your que. If you ask for code then link is in description.

[@debkumar5536]

Sir I need the code for the model. In link you provided is for model and I unable to get the code from that

[@debkumar5536]

I wants to see how you write the code for vgg model

[@karanpatel3358]

Deb Kumar I wrote just VGG16 model’s dense layer you can find the code in any VGG16 models. If you have still doubt contact me on LinkedIn.

[@ankitupadhyay3546]

I want to know more about this model can you help me out

[@karanpatel3358]

Yeah sure, Tell me what you want to know ?

[@ankitupadhyay3546]

where i can contact you

[@ankitupadhyay3546]

I want to know about that epoc model

[@ankitupadhyay3546]

And on what basis and how detection is done

[@karanpatel3358]

@@ankitupadhyay3546 We can communicate on LinkedIn.

[@SGandhi]

I want to understand full process with code in python can you help me with my below question

[@RS-Rangoli]

How to import download data set to google collab

[@karanpatel3358]

rupasri ch it’s widely depends on datasets and platform which you are using for code. For example Goggle Colab has the facility to read data from drive. I hope you clear with your doubts.

[@abdelkaderwalidmekibes9469]

Hi, thank u for the video, i have question, in the non-covid class, there is images which contain white zones, is it that covid-19?

[@karanpatel3358]

No it may be because of its "pneumonia patient" or lung damange other than COVID.

[@abdelkaderwalidmekibes9469]

@@karanpatel3358 thank you for ur answer

[@pallavipriyadarshini5065]

Very nice explanation.. Can u plz send me ur github link for this code..

[@karanpatel3358]

Sure please write “ karan 1210 GitHub “ on google . Or click on link karanpatelbme.blogspot.com/?m=0 ( here you find GitHub Icon click on that )

[@pallavipriyadarshini5065]

@@karanpatel3358 thank u

[@pramodkumarsahoo3351]

@@karanpatel3358 please suggest i am unable to load ur vgg model for 50 epoch..

[@karanpatel3358]

@@pramodkumarsahoo3351 write ✍️ your error. Because it’s working in my system.

[@অনন্যা-ষ৫ফ]

Hello sir. How can I use your model in my code? I am working on a google collab. I am a little bit confused. I will be really grateful if you give me the solution.

[@karanpatel3358]

Probably you can download the file extension and then load it to your model.

[@অনন্যা-ষ৫ফ]

@@karanpatel3358 how can I contact you? I need some help.. could you please help me?

[@অনন্যা-ষ৫ফ]

could you please provide me the model code. I mean training parts code. model fitting code? the h5 file is not working in my code. please please please , it will be really helpful for me. please sir.

[@karanpatel3358]

@@অনন্যা-ষ৫ফ Linkedin OR write karan patel biomedical on google

[@অনন্যা-ষ৫ফ]

@@karanpatel3358 Ok thanks

[@AnkitKumar-jb5qf]

After entering the model you provided, in the output, the images are getting black, please help .. btw great video very useful

[@karanpatel3358]

its not like image getting black i guess, because there is no line of code which actually do that. But its may be because you choose input image like that.

[@AnkitKumar-jb5qf]

@@karanpatel3358 i choosed the same dataset that you provided

[@karanpatel3358]

@@AnkitKumar-jb5qf then please tell me what kind of drakness you are talking about ? because it might possible that image it self from input side is also that much dark. Thank you.

[@AnkitKumar-jb5qf]

@@karanpatel3358 can I share my no..? Btw .. where I right plt.imshow(X_test[index]) plt.show() After this ..

[@karanpatel3358]

@@AnkitKumar-jb5qf we will Communicate on LinkedIn

[@mohammedal-rubaie8902]

how i can create vgg dataset for my own data like yours please help

[@karanpatel3358]

Please Watch video " Detect COVID19 from chest X-ray || Deep Learning algorithm to detect corona from chest X-ray images." In this video I describe you the technique.

[@pallavipriyadarshini5065]

Can we use the same code of vgg model for covid chest x ray and covid ct scan both in same way for all cnn models

[@karanpatel3358]

@@pallavipriyadarshini5065 Good Question but you can not use the same model because, it had train on CT scan images.

[@vinaykumar-xz6eh]

hi that vgg_ct_trail_50_epoch_karan_patel.h5 link is not working can u please send the link in the comments

[@SGandhi]

Hoe we can segment the image of covid , pneumonia patient in python can you just make vidoe fir that and how can we do feature selection also

[@qya.4594]

Would you please remove the annoying musics?

[@karanpatel3358]

I definitely keep this note in future videos.

[@inesaoues5464]

I sent you an e-mail

[@karanpatel3358]

I’ll respond to that Thank you 😊