Dissertation Talk: Phase-space imaging in computational imaging

Seminar | May 2 | 11:30 a.m.-12:30 p.m. | Soda Hall, VCL center, 510

 Hsiou-Yuan Liu

 Berkeley Center for Computational Imaging

Computational imaging has open many possibility in the study of imaging; examples are digital refocusing, phase retrieval and diffuser cameras. The optics in use does not necessarily form an image of the target, as done in traditional imaging, but capture essential features for a post-capture computation to reconstruct the image. The design of computational part will also affect how the optics is built so computational imaging is a joint hardware-software design process. It can achieve, e.g., spacing constraints of some application or removing the motion hardware in large-field high-resolution microscopy. More than just reconstructing images, computational imaging also allows us to capture physical quantities that can only be captured indirectly, such as phase space of light.
In this talk, I will discuss using computational imaging to capture the phase space of light and the latter's applications on 3D imaging and imaging through scattering. Phase space captures additional angular (spatial frequency) distribution to traditional images and thus is useful for applications that put information into angles, such as light traveling along its angle or scatterer smear a light rays into multiple angles. At last, we will point out that phase space can contribute back to computational imaging by helping the joint design process.