Prashant discusses the limitations of current research that primarily focuses on down-sampled images, which often leads to a loss of crucial information. He highlights the innovative patch-based approach his team employs to analyze full-resolution images, revealing that traditional models struggle to converge when applied to larger sizes. Despite the challenges, he notes a gap in the literature regarding the testing of larger images, suggesting a unique opportunity for advancement in the field.
