Hi, Followed your link from my Twitter feed and glad I did. Your tutorial was enjoyable. Your depth of knowledge is impressive and your code clean and well documented. Liked and Subscribed. Thanks. Question: Is it possible to converting it down to Int8 for microcontrollers?
Hi Thanks for the video. Very detailed. I could not help but subscribe. Should i want to keep the color of the skin lesions in the segmented image how can i do that? Thanks.
@@ranjani634 I used a different code than this video but I basically isolated the hsv channels of image, gave a colour range whereby the segmentation would only affect values outside the range and converted area of the image whose pixels have been whitened back to their original colour. I'll try and find the TH-cam video that I used but you'll have to tweak the code a bit
Good tutorial with detailed explanations. I really enjoyed watching this video but cannot use it for categorical data set even after changing 1 to 7 and sigmoid to softmax. It is complaining of incompatible shapes. please help me out
HI, Thank you so much for this useful video.Could you please do a video for a scenario like when there is images but all are not unhealthy.Groudtruth/mask is available only for diseased ones. To do a segmentation for these kind of issue
Hello my friend We have to make a memory on the skin lision segmentation and we need your help We are not getting the results The results file is empty And we get the metrics without the results saved
@@IdiotDeveloper yes i know but i am unable to find the masked images of iris in dataset that is why i am requesting you to please find the dataset that suits you and please make a tutorial on the( human)iris segmentation through U-NET thanks in advance
Thank you soo much, but I am getting ModuleNotFoundError at the step “from model import build_unet” No module named ‘model’ I am doing this code on google colab Please can you tell how to fix this error
Sir kindly share a vedio about brain tumour segmentation by using CNN with U-Net architecture. I humbly requested to you. Plz in that vedio explain code step by step. And also share the link of dataset. I shall be very thankful to you.
Precision is ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior. How can l correct this occurs when running eval.py
Sir I am working in this code for the skin cancer segmentation. I have write the same code which you have uploaded in this vedio. But when I run the code to split the dataset into train, validation and test dataset. So it's give an error. The error is " test_size =0 should be either positive and smaller than the number of samples 1 or a float in the (0, 1) range". Kindly guide me what I will do for this. I shall be very thankful to you for this kindness. Regards.
thanks for the video.. while running the eval.py file i got ValueError: test_size=0 should be either positive and smaller than the number of samples 0 or a float in the (0, 1) range. can u pls help on it... thanks in advance
then i used for 1008 images it splits into 606:201:201 still i m getting same error while loading the test data can u pls help me to rectify this..... thanks in advance
@@padmavathiv2429 this error is occur due the image and masks path. when you give the images and masks path so in this path should be write like this. images = sorted(glob(os.path.join(path, "/content/drive/MyDrive/skin_cancer_2/images/*"))) masks = sorted(glob(os.path.join(path, "/content/drive/MyDrive/skin_cancer_2/masks/*"))) you can see the last /* sign. I have got the same error. when I changed it so its reduced. Thanks.
Hi,
Followed your link from my Twitter feed and glad I did. Your tutorial was enjoyable. Your depth of knowledge is impressive and your code clean and well documented. Liked and Subscribed. Thanks.
Question: Is it possible to converting it down to Int8 for microcontrollers?
Thanks 🙏 for following.
I don't know about converting it to int8.
thank you so much !! cannot thank you enough, how about Task 2: Lesion Attribute Detection and Task 3: Disease Classification? will you make a series?
That's a good idea. I will think upon it.
thank you so much sir for the video...
Hi Thanks for the video. Very detailed. I could not help but subscribe. Should i want to keep the color of the skin lesions in the segmented image how can i do that? Thanks.
did you get the answer?
@@ranjani634 I used a different code than this video but I basically isolated the hsv channels of image, gave a colour range whereby the segmentation would only affect values outside the range and converted area of the image whose pixels have been whitened back to their original colour. I'll try and find the TH-cam video that I used but you'll have to tweak the code a bit
thanks for this video.
Good tutorial with detailed explanations. I really enjoyed watching this video but cannot use it for categorical data set even after changing 1 to 7 and sigmoid to softmax. It is complaining of incompatible shapes. please help me out
HI, Did you found a solution for categorical data set
HI, Thank you so much for this useful video.Could you please do a video for a scenario like when there is images but all are not unhealthy.Groudtruth/mask is available only for diseased ones. To do a segmentation for these kind of issue
Hello my friend
We have to make a memory on the skin lision segmentation and we need your help
We are not getting the results
The results file is empty
And we get the metrics without the results saved
hello, I tried out your code. but when I load the model and tried to train it for few more epochs with model.fit, Error says " TypeError: '
Iam also this model it predict only cancer image,if there is no cancer in that image how will predict them
please make a video on iris segmentation through u net
Sure, I will make a video on it. Can you provide me with the link to the dataset?
Do share and SUBSCRIBE. 🙏
For segmentation we need mask also along with input images.
@@IdiotDeveloper yes i know but i am unable to find the masked images of iris in dataset that is why i am requesting you to please find the dataset that suits you and please make a tutorial on the( human)iris segmentation through U-NET thanks in advance
Thank you soo much, but I am getting ModuleNotFoundError at the step “from model import build_unet”
No module named ‘model’
I am doing this code on google colab
Please can you tell how to fix this error
hey, my eval.py file is working but my result images are not showing, there is no error in code, i have checked it, please guide
Hi
I had used your code for ALL_IDB1 dataset,but unable to display predicted mask.its showing blank,how can I resolve this
i don't find the data used in this video ???
how to find the area of lesions using U-net can you suggest..
Can you make videos about Image registration? Thanks.
Sir kindly share a vedio about brain tumour segmentation by using CNN with U-Net architecture. I humbly requested to you. Plz in that vedio explain code step by step. And also share the link of dataset. I shall be very thankful to you.
Thanks for your valuable feedback. I surely work upon brain tumor segmentation.
@@IdiotDeveloper sir i am waiting for that vedio. Kindly use MR images dataset for brain tumour segmentation.
Regards.
@@IdiotDeveloper can you help working on kits19 kidney tumor segmenattaion
Precision is ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior. How can l correct this occurs when running eval.py
also the same problem
Sir I am working in this code for the skin cancer segmentation. I have write the same code which you have uploaded in this vedio. But when I run the code to split the dataset into train, validation and test dataset. So it's give an error. The error is
" test_size =0 should be either positive and smaller than the number of samples 1 or a float in the (0, 1) range".
Kindly guide me what I will do for this.
I shall be very thankful to you for this kindness.
Regards.
There might be some issue with the path of the dataset.
So what i will do for this?. I mean how I do changes?
Can you make a video on weeds segmentation from background.
thanks for the video..
while running the eval.py file i got ValueError: test_size=0 should be either positive and smaller than the number of samples 0 or a float in the (0, 1) range.
can u pls help on it...
thanks in advance
I think there is some issues with the images or masks path. Check whether the list contains some path or not.
@@IdiotDeveloper i used for 100 images totally
it splits into60:20:20
then i used for 1008 images
it splits into 606:201:201
still i m getting same error while loading the test data
can u pls help me to rectify this.....
thanks in advance
@@padmavathiv2429
this error is occur due the image and masks path. when you give the images and masks path so in this path should be write like this.
images = sorted(glob(os.path.join(path, "/content/drive/MyDrive/skin_cancer_2/images/*")))
masks = sorted(glob(os.path.join(path, "/content/drive/MyDrive/skin_cancer_2/masks/*")))
you can see the last /* sign. I have got the same error. when I changed it so its reduced.
Thanks.
Please do heart segmentation on ct scan image
I currently do 2d segmentation. I think CT scan images are 3d.
@@IdiotDeveloper yes
can you make video on 2d segmentation video also
Hey how many years it take to code you in such a way
It's an ongoing process, it improves with time.