Models camera poses as probabilistic distributions with learnable uncertainty to enable robust surface reconstruction under significant pose errors.

Performs Global-to-local analysis using Progressive learning guided by uncertainty quantification to identify regions of diagnostic ambiguity followed by detailed evidence based classification.

Combine Audio and Video Modalities utilizing graph based techniques to identify deception in court-trials and game shows.

Utilize Few-Shot learning techniques to identify diseases with the help of plant leaf images.

Entropy-guided label pruning and region-aware uncertainty estimation enables fracture diagnosis models to reason under ambiguity.

Deep learning-driven MRI analysis improves brain tumor detection, segmentation, and grading by leveraging advanced feature extraction, attention mechanisms, and ensemble classification for precise diagnosis.

A novel deep learning approach enhances Alzheimer's detection from MRI scans by integrating multi-scale feature extraction and a Bi-Focal Perspective mechanism to highlight critical pathological markers.

A deep learning framework enhances breast cancer diagnosis from ultrasound images using precision mapping for accurate lesion segmentation and multi-level attention for improved classification.

An Ensemble model with weight gradient optimization function to improve plant leaf disease detection.