Ambient light in turbid media (water, biological tissue, fog) is strongly attenuated in turbid media, making it difficult to illuminate a subject naturally. This calls for artificial illumination to compensate for this effect. This study acknowledges this necessity as well as the strong degradation of signal-to-noise ratio (SNR) due to backscatter which is enhanced by artificial illumination. Most prior methods for visibility improvement use active illumination scanners (structured and gated), which are slow and cumbersome. We propose multiple approaches to model images under these effects and to use artificial illumination to our advantage to perform image recovery.
We show that by imaging with polarized light in two image frames, backscatter can be removed. Also, that by fusing a few image frames under illumination from varied directions, backscatter can be minimized and the scene can be illuminated evenly without shadows. Finally, we show that by using fluorescence imaging techniques, it is possible to eliminate the backscatter component and increase the SNR.
Publications
In Situ Target-Less Calibration of Turbid Media. Ori Spier, et al. ICCP, 2017.
Active Polarization Descattering. Tali Treibitz and Yoav Y. Schechner. IEEE Trans. Pattern Analysis & Machine Intelligence, 2009.
Photometric Stereo in a Scattering Medium. Zachary Murez, et al. ICCV, 2015.
Instant 3Descatter. Tali Treibitz and Yoav Y. Schechner. IEEE CVPR, 2006.
Turbid Scene Enhancement Using Multi-Directional Illumination Fusion. Tali Treibitz and Yoav Y. Schechner. IEEE Trans. on Image Processing, 2012.