ASL Sign Language Detection Using CNN | Deep Learning | Python | Tensorflow
HTML-код
- Опубликовано: 23 дек 2023
- ASL Sign Language Detection Using CNN | Deep Learning | Python | Tensorflow
In this video, I'm giving a tutorial on making a sign language detection model using CNN( Convolution Neural Network) (Deep Learning neural network architecture) in Python and Tensorflow.
In this, I'm using Google Colab to train the CNN model.
===========================================================
Tool Used:
Python - 3.10.7
Visual Studio Code( Vs code)
Google Colab
==========================================================
Library/Modules/Packages used:
opencv (pip install opencv-contrib-python)
tensorflow (pip install tensorflow)
keras (pip install keras)
splitfolders (pip install split-folders)
==========================================================
Dataset Used:
github.com/kum...
==========================================================
Source Code:
github.com/kum...
=========================================================
Follow me on:
Instagram - / progresswithpython
LinkedIn - / vivek-kumar-689761231
Like, Share, Comment and Subscribe :)
Really helpful I will get full marks because of you, thank you so much
lesssgooooo maanuuuuuu we made itttt!!!!!
gareebo mai mithai batwao bhai
I am starting it can u help me if I get some errors or smthg?
Vai dataset kaise create kiye ho?
si r yek project real time sign detection and text recognition per yek video banaye , or usme hello , good morning ,good night and other so many sing add kare plz sir
Brother you might just save my grade here, I made a model but it's not working as well so hopefully following yours will help
sir sign language to text and audio project video chaiye using cnn
good work brother🥰🥰🥰
Where did he bring the file "Untitled-1.py" ? where can i find it ?
very good work. It really helped me a lot.
Thanks
logits and labels must be broadcastable: logits_size=[128,6] labels_size=[128,26]
[[{{node categorical_crossentropy/softmax_cross_entropy_with_logits}}]] [Op:__inference_train_function_6550] What to change now?
Change 6 to 26 in model layers
hey we are doing this for our BE project for final year and i had some doubts..
1. we are working with indian sign language. will this model be able to work for all alphabets(for around 26000 images in total) and will it work for words too?
2. if we want to give front end to this will it be possible and what do we use?
3. we will also add text to speech conversion so will that be possible with this output? and also if we can spell out words alphabet by alphabet.
sorry for so many doubts.. thanks for the video!
can I ask where you got the dataset for Indian sign language
Were you able to make the project? Or find a good reference?
You can use tkinter library of python for doing this.
I am also working on this project but it is not working if background changes.
@@vagisangovindan7123 we are creating our own dataset but we are creating it in black n white acc to this video! but there are more available on kaggle
@@cosmic_hz well we are creating it now by using this video as a base and will be adding more pre processing steps
I m getting the error of Class name sequential could not locate....Could you please tell the solution for it...i think it's having some issues in the json file
this error is coming with the latest tensorflow version.
and its working with these version:-
tensorflow==2.10.0
opencv-contrib-python==4.6.0.66
@@ProgresswithPythonwhere can i download this version of tensorflow? When i try with pip i get the following error: Could not find a version that satisfies the requirement tensorflow==2.10.0
@@ilayyy.dddaaa I'm using Python 3.10.7 and tensorflow of version 2.10.0 is installing in Python 3.10.7
sir es per bss A,B,C,D detect karta hai ky
Hello,can you please help us,we are doing project on British sign language,but facing so much of issues...
When i split my folder .folder is splitted but files are empty any one knows about the solution
Bro please upload flow of CNN architecture in image
Man how to add our own data an d also words detection as well?
bhai sign to text conversion ki ek video banao
How to download png images into my folder
Hii sir can please share the source and also setup the project
You didn't provide json file at git code
Sir dataset kaise create karenge?
But it is not giving as accurate results as u were showing. It is showing either S or M. Are we doing anything wrong?
Please reply
Make sure your dataset is proper and make dataset large with variations hand poisition, angle
@@ProgresswithPython we used the same dataset that u used from your GitHub
But we used your dataset which was in your github link which is the same dataset that u used
@divyaaaa5420 you have to add your own dataset in my dataset so that the model will able to learn that hand background is not a feature.
I have to complete this project by tomorrow, hope to see your reply soon
hlo is ur project complete i want it
I have followed all steps but at last while running realtimedetection.py it is showing error that Could not locate class 'Sequential
yess... did you find any way to correct that error
Yes same error any idea how to correct it?
@@ravalimanthena8935 try installing tensorflow and keras 2.10.0 version my webcam is working atleast but accuracy is fluctuating
@@ravalimanthena8935 did you find the solution as i m also getting the same error.
@@jaipriyam4977 did you find the solution as i m also getting the same error.
Sir dataset kaise add karu?
Which version of Keras you are using??
keras - 2.10.0
tensorflow - 2.10.0
python - 3.10.7
google collab link buddy ?
share it ..
Link of repo in the description, which contains all code files including colab file
What’s the accuracy
Bro tumne CNN model mei classname kaise define ki like agar mai 6 se zyada classes chaahu toh kya karu?
Output layer mai 6 ki Jagah jitne classes hai Utne likh do
@@ProgresswithPython Ok bro done.Thanku
Bro is model ki accuracy nikalne ka aur graph banane ka code bata do plz report mei daalna hai
@@arpanbajpai1219 ruclips.net/video/SpnBXS-XzRg/видео.htmlsi=FbrpLKgPgMjQOljv
Watch this for model accuracy and for confusion Matrix graph
You have not given colab file in github
There is a file with the name trainmodel.ipynb in github repo
@@ProgresswithPython How to open it
BUDDY SEQUENTIAL ISSUE AA RAHA HAI KYA KARU
this error is coming with the latest tensorflow version.
and its working with these version:-
tensorflow==2.10.0
opencv-contrib-python==4.6.0.66
sir collab ka link dedo pls
mil nhi raha hai
aapne nhi diya
github.com/kumarvivek9088/SignLanguageDetectionUsingCNN/blob/main/trainmodel.ipynb
bhaiya ek help krdo
mughe yeh project bnana tha from A to z aap pls pura project bna do in YT @@ProgresswithPython
ruclips.net/video/1FOawpuQflg/видео.html
@@debarghyakundu908 did u make?
signlanguagedetection48x48.json file ??
Present in the link of source code
@@ProgresswithPythonthat link didn't work now
@@Riddles_World548 ok I will provide the model link but my model is trained on those 6 classes only which I showed in the video
@@ProgresswithPython I need that only
@@Riddles_World548 well you can run that train model file, it will take hardly 10 to 15 min and the model will get saved to your google drive
Hello, which versions of opencv,tensorflow and keras
keras - 2.10.0
tensorflow - 2.10.0
python - 3.10.7
opencv - 4.6.0.66
If i get your mail id,i will mail the folder of tge dataset collected,please help
Upload the dataset folder to your google Drive and give access to that folder to me. Or upload the dataset to github repo
@@ProgresswithPythonokay
@@ProgresswithPythondrive.google.com/drive/folders/1KjaqYIInF_fsQ1u3_KrshaAKCCCnJ6Jf?usp=sharing
This is the link to the dataset collected
Please reply !
I have followed all steps but at last while running realtimedetection.py it is showing error that Could not locate class 'Sequential
same for us
I think because different version tensorflow
i am also getting this error do you solved it? if yes then can you plz tell me how
Got the same error message, did anybody get to correct it?
@@karthikeyanr6321 did anybody get to correct it, i am also finding same error.