Research

Developing a vision-language model that connects text descriptions with ultrasound images, facilitating standard plane navigation and reducing variance among radiologists.

Developed a system for detecting abusive texts in online content, focusing on IPV research, using a tailored deep learning model and an interactive dashboard for Nepali languages.

Developed a multi-modal vision-language approach for accurate segmentation of medical images by extracting semantic information from image descriptions, evaluating and exploring the impact of variations in descriptions on model performance.