Nima Khademi Kalantari, Ph.D. Candidate, University of California, Santa Barbara
The illumination of the world we live in has high dynamic range, and we must therefore capture it properly in order to reproduce environments as our eyes perceive them. Unfortunately, standard digital cameras have low dynamic range and can measure only a small range of the total illumination of the scene. Recently, high dynamic range (HDR) imaging has received great attention because of its tremendous potential in transforming the world of photography. Approaches have been proposed to generate high-quality HDR images and videos using specialized camera systems. However, because of their high cost and general unavailability, these cameras are impractical for the average consumer.
To make high-quality HDR imaging widespread, we must focus on approaches that use standard digital cameras. The most common approach is to take sequential LDR images at different exposure levels (known as bracketed exposures) and then merge them into an HDR image. Although this technique can produce spectacular results, the original approaches work only for static scenes because they typically assume a constant radiance at each pixel over all exposures. For a dynamic scene, this method produces ghost-like artifacts from even small misalignments between exposures. In this talk, I will present our research on this topic which resulted in two methods for generating high-quality HDR images and videos from bracketed exposures. Our methods work by minimizing a patch-based objective function to correct for the motion between different bracketed exposures in order to generate ghost-free results.
About Nima Khademi Kalantari:
Nima Khademi Kalantari is a Ph.D. student at UCSB working under the supervision of Dr. Pradeep Sen in the MIRAGE Lab. His research interests focus on the application of signal processing to computer graphics and vision. He has published 9 journal papers in different areas including sampling, rendering, and imaging.
Hosted by: UCSB Center for Visual Computing