Paras Pant

Papers

Total Citations

H-Index

About

Paras Pant’s research lies at the intersection of spectral imaging, computer vision, and computational optics, with a focus on disentangling complex light interactions in digital imagery. His most-cited work, “Highlight detection and removal from spectral image” (2011, 18 citations), introduces a constrained spectral unmixing method that separates diffuse and highlight reflections from a single spectral image. By imposing positivity and sum-to-one constraints on reflection fractions, Pant’s approach enables accurate removal of specular highlights—a persistent challenge in material recognition, remote sensing, and medical imaging. This contribution is particularly notable for its elegance in requiring only one input image, avoiding the need for multiple exposures or polarizing filters. Beyond this flagship paper, Pant has explored topics in hyperspectral data analysis and image enhancement, demonstrating a consistent ability to solve practical problems in spectral processing. His work has been cited by researchers in optics, computer graphics, and machine learning, reflecting its interdisciplinary reach. For students and researchers, Pant’s methodology offers a clear, reproducible framework for tackling highlight artifacts, making his research a valuable reference in the growing field of computational spectral imaging.