American Sign Language Detection

[aditya0929]

Pull Request for DL-Simplified 💡

Issue Title : American Sign Language Detection #312

• Info about the related issue (Aim of the project) : to correctly predict the sign language images given in the dataset
• Name: ADITYA NARAYAN JHA
• GitHub ID: https://github.com/aditya0929
• Email ID: [email protected]
• Idenitfy yourself: (Mention in which program you are contributing in. Eg. For a JWOC 2022 participant it's, JWOC Participant) SSOC-2023 Participant

Closes: #312

Describe the add-ons or changes you've made 📃

I chose convolutional neural networks (CNNs) for the multi-instance classification situation since the dataset consisted of 36 different classes of sign language from (A-Z) to (0-9), and each class needed to be accurately identified to improve the overall accuracy of the model.

I decided to utilize four pre-trained models on the ImageNet dataset: VGG16, InceptionResNetV2, InceptionV3, and MobileNet.

VGG16 is a deep CNN with 16 layers and is well-known for its exceptional performance in image recognition tasks. It has been widely used as a benchmark model in computer vision research.

InceptionResNetV2 combines the concepts of the Inception and ResNet models. It is a powerful CNN architecture with improved accuracy and efficiency.

InceptionV3 is another CNN architecture that has been successful in image classification tasks. It incorporates inception modules to capture multi-scale features.

MobileNet is specifically designed for mobile and embedded devices with limited resources. It offers a good balance between accuracy and computational efficiency.

Each of these models has its own strengths and is suitable for different scenarios depending on the available resources and desired trade-offs between accuracy and computational requirements.

In the "models" folder, I have included all the models within separate ipynb files for all the models that i have made .

Type of change ☑️

What sort of change have you made:

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Code style update (formatting, local variables)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested? ⚙️

i tested it by making a classification report after fitting the models to the certain epochs required and then copying a file path to predict_images( ) and displaying the correct prediction and each time it matched the class name of the image directory
this is the accuracy comparison of different models

Accuracy Comparison

Model	Accuracy
VGG16	97%
InceptionV3	88%
InceptionResNetV2	88%
MobileNet	97%

these are the predicted labels

Checklist: ☑️

• My code follows the guidelines of this project.
• I have performed a self-review of my own code.
• I have commented my code, particularly wherever it was hard to understand.
• I have made corresponding changes to the documentation.
• My changes generate no new warnings.
• I have added things that prove my fix is effective or that my feature works.
• Any dependent changes have been merged and published in downstream modules.

[github-actions]

Our team will soon review your PR. Thanks @aditya0929 :)

[abhisheks008]

Your PR is approved and ready to be merged.
@aditya0929