Utilised complementary strengths of foundational models to achieve state-of-the-art performance in novel object detection.
VANE-Bench is a benchmark designed to assess Video-LMMs’ ability to detect and localize anomalies in videos, revealing their limitations in subtle anomaly detection through a visual question-answering challenge.
Developed an complete end-to-end system for using forehead creases as a biometric modality for human authentication. Code and Database have also been made public.
Machine Learning models were developed and trained to classify and detect malicious URLs.