In recent years, computer vision has seen great advances in the areas of object detection, recognition, segmentation and tracking, 3D reconstruction, autonomous navigation, image retrieval, and industrial inspection. These advances have been made possible due to successes in the core areas of scene sensing and interpretation. However, almost all of this research is still based upon one fundamental assumption --- that light reflected by a surface reaches the sensor unaltered. For the large part of the past 35-40 years, image formation has been defined as a geometric mapping from 3D to 2D, which inherently leads to loss of information. We strongly argue that light transport in scattering media must not be viewed as noise that a traditional vision algorithm needs to overcome, but rather as a new form of encoding of light and hence, the images themselves. Our objective in this symposium is to increase awareness about this research area.

The past decade has seen computer vision research double in size in terms of both the number of papers in journals and conferences and applications. The field has converged in new ways with other fields like machine learning, graphics and medical imaging. As a result, there is a fast growing demand for vision techniques in various scientific fields ranging from oceanography (underwater imaging) to astronomy (telescope and satellite imaging) to remote sensing (aerial imaging) and to even biology and medicine (microscopy, endoscopy, tomography). In all these areas, however, there is no escape from light scattering. We believe this symposium can inspire research in this area increasing the impact of vision in many application domains.

We also believe that the symposium will spur new interdisciplinary collaborations among researchers in diverse fields. The invited talks in this symposium will help focus on both the remarkable similarities in light transport research problems faced in a range of disciplines, as well as their distinctive aspects.