Publications

This paper presents a comprehensive approach to tuberculosis detection from chest X-ray images using vision transformers and transfer learning. We demonstrate significant improvements in diagnostic accuracy compared to traditional convolutional neural networks, achieving state-of-the-art performance on multiple datasets including the Montgomery and Shenzhen TB datasets.

We propose an automated system for COVID-19 detection using both chest X-ray and CT images. Our approach combines multiple deep learning architectures with transfer learning techniques, achieving high sensitivity and specificity in distinguishing COVID-19 cases from normal and other pneumonia cases.

This research introduces a novel approach combining edge detection techniques with graph neural networks for ECG signal classification. Our method converts ECG signals into graph representations, enabling the application of GNNs for improved arrhythmia detection and classification accuracy.

We present BLOOM-LoRA, a parameter-efficient fine-tuning approach for adapting the BLOOM large language model to medical applications. Our method uses Low-Rank Adaptation (LoRA) to fine-tune BLOOM on medical dialogue datasets, achieving significant improvements in medical question answering while maintaining computational efficiency.