SayantikaLaskar commented 6 days ago

Pull Request for DL-Simplified 💡

Issue #771

Issue Title : kidney stone detection

Info about the related issue (Aim of the project) : The goal of this project is to develop the performance of three distinct deep learning architectures—VGG-like, CNN with spatial attention, and ResNet-like—for image classification tasks. By training these models on a dataset containing approximately 12,000 images across four classes, the project aims to evaluate and analyze their respective accuracies, loss values, and generalization capabilities. This comparative study will not only highlight the strengths and weaknesses of each architecture but also provide insights into the effectiveness of incorporating spatial attention and residual connections in deep neural networks. Ultimately, the findings from this project will guide the selection of the most suitable model architecture for robust image classification in practical applications.
Name: Sayantika Laskar
GitHub ID: (https://github.com/SayantikaLaskar)
Email ID: sayantikalaskar2002@gmail.com
Idenitfy yourself: GSSoC 2024 participant

Closes: #771

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

This project aims to develop and compare the performance of three advanced deep learning architectures—VGG-like, CNN with spatial attention, and ResNet-like—for classifying images into four categories: Normal, Cyst, Tumor, and Stone. The dataset, consisting of about 12,000 images, is split into training, validation, and test sets. Each model is trained with augmented data to improve generalization. The models are optimized and their performances are evaluated based on metrics like accuracy, precision, recall, and F1-score. The goal is to determine which architecture best balances accuracy and generalization, offering insights into the benefits of spatial attention and residual connections in deep learning for medical imaging.

Type of change ☑

What sort of change have you made:

[x] Bug fix (non-breaking change which fixes an issue)
[x] New feature (non-breaking change which adds functionality)
[x] 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? ⚙

The project involved evaluating three deep learning architectures—VGG-like, CNN with spatial attention, and ResNet-like—for classifying kidney condition images into four classes (Normal, Cyst, Tumor, and Stone). The dataset, consisting of 12,000 images, was split into training, validation, and test sets. Training data underwent extensive augmentation to enhance model generalization. Each model was trained with categorical cross-entropy loss and Adam optimizer, using callbacks for early stopping, best model checkpointing, and learning rate reduction. Performance was assessed on a test set using accuracy, precision, recall, F1-score, and confusion matrices, with results visualized through accuracy curves and detailed classification reports. This comprehensive evaluation aimed to identify the most effective architecture for accurate kidney condition classification.

Checklist: ☑

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