Research and engineering projects from Srijan:CoE.
Leveraging generative AI in healthcare for disease progression assessment, medical image super-resolution, medical image synthesis from reports (MPD-CXR), and non-invasive disease detection (e.g., TB using AcouSem-AFNet, Silicosis).
activeFocusing on computer vision improvements for adverse weather, preparing the FogRoad-12M dataset, and developing dehazing and deraining algorithms.
activeEnhancing teaching-learning through generative AI with tools for personalized educational examples (ExaCraft), automated handwritten answer sheet evaluation (AgiQ-A), complex table visualization (TableSense), and symbol table extraction.
activeIntroducing attack frameworks and guided diffusion-based models for removing, detecting, and distinguishing invisible watermarks in synthetic media.
activeDeveloping frameworks and benchmarks (like Genµ) for responsible generative AI to selectively forget targeted concepts while preserving unrelated knowledge.
activeEvaluating and quantifying stereotypes within MLLMs and text-to-image generative models, and developing multilingual approaches to improve fairness across languages.
active