The world around us has an immense range of light levels from bright white objects under direct sunlight to dark objects in the shadows under moonlight. Our visual systems are capable of quickly and effortlessly adapting and adjusting to these changes in illumination level to allow useful perception of objects both in the shadows and in bright light (and sometimes both within a single scene). Most cameras do not have similar capabilities, but rather must be set to properly expose only a limited range of objects in a scene. Thus, while our visual system rarely over- or under-exposes parts of a scene, cameras often do and thus limit the range of objects that can be seen in the image. How does the human visual system succeed in handling the immense dynamic range of our visual world?
With support of the National Science Foundation and under the supervision of Dr. Fairchild, Jiangtao Kuang will study the perception of high dynamic range scenes and images, and develop mathematical techniques needed to manipulate and display these images. This dissertation improvement award will be used for the initial construction and testing of a camera and measurement system to provide better images and information about those images for ongoing research on scene and image perception. The result will be a camera system capable of capturing high dynamic range images in the field, a color measurement system capable of accurately measuring the physical colors of objects in the original scene, and a perceptual scaling system allowing the perceived color of objects in the original scene to be carefully recorded. A database of images with these three sets of information (image, physical color measurement, perceived color) will be made available for future research on perception and imaging. This database will contribute to a wide range of research efforts related to the function of the human visual system.
