Pinned Loading

1. CryoET-Protein-Complex-Annotation-Classification CryoET-Protein-Complex-Annotation-Classification Public

   This project processes electron tomography (ET) data to detect and segment biological particles like proteins and viruses. It uses Copick for particle detection and applies segmentation algorithms,…

   Jupyter Notebook

2. DermaStratif-Multiclass-Lesion-Stratification-and-diagnosis DermaStratif-Multiclass-Lesion-Stratification-and-diagnosis Public

   DermaStratif classifies skin lesion images into 8 disease categories using EfficientNet-B0 with advanced fine-tuning techniques like Base Training, Adapter Fine-Tuning, and LoRA Fine-Tuning. The pr…

   Jupyter Notebook

3. Adversarial-ResNet-Resilience-Gradient-Perturbation-Fidelity-Analysis- Adversarial-ResNet-Resilience-Gradient-Perturbation-Fidelity-Analysis- Public

   This project explores generating adversarial examples and fine-tuning the ResNet18 model on the MNIST dataset using techniques like FGSM and PGD. The model achieved 97.91% accuracy on clean data bu…

   Jupyter Notebook

4. Universal-Prompt-Injection-for-LLMs-Adversarial-Evasion-Response-Manipulation- Universal-Prompt-Injection-for-LLMs-Adversarial-Evasion-Response-Manipulation- Public

   This project demonstrates how prompt injection can manipulate language models (LLMs) like GPT to generate targeted responses, with a 100% success rate in influencing the model's output.

   Jupyter Notebook

5. Voice-to-Text-Summarizer-via-Transfer-learning-and-Semantic-Representation Voice-to-Text-Summarizer-via-Transfer-learning-and-Semantic-Representation Public

   This project builds a Voice-to-Text Summarizer using Hubert and Wave2Vec for speech recognition, and BART and BERT for summarization. Hubert outperformed Wave2Vec with 90.26% accuracy. BART and BER…

   Jupyter Notebook

6. Multi-Modal-multi-class-Osteoperosis-Osteopenia-Diagnosis-with-Hybrid-Fusion Multi-Modal-multi-class-Osteoperosis-Osteopenia-Diagnosis-with-Hybrid-Fusion Public

   Advanced AI system combining medical imaging and clinical data for accurate osteoporosis and osteopenia classification using state-of-the-art multi-modal fusion techniques.

   Jupyter Notebook