Coded Computational Imaging

 

CVPR 2010 Tutorial
Monday, June 14, 2010  (Half Day, AM Session)
Amit Agrawal, Ashok Veeraraghavan, Srinivasa Narasimhan and Ankit Mohan



Computational imaging offers imaging functionalities beyond traditional imaging. Recently, we have seen tremendous interest in new forms of sampling the visual world. These include capturing the angular information in a light field camera for view interpolation and digital refocusing, modulating the temporal integration pattern of a camera for motion deblurring via coded exposure, modulating the aperture pattern for depth estimation and digital refocusing etc. Coupled with powerful reconstruction algorithms, computational imaging simplifies several vision problems.


In this course, we will present an extensive overview of computational imaging techniques while providing key ideas and insights into their workings. The participants will learn about topics related to computer vision, computational photography and computer graphics. We hope to provide enough fundamentals to satisfy the technical specialist as well as tools/softwares to aid graphics and vision researchers, including graduate students.




Course Description


Introduction (Srinivas, 10 mins)     pdf

    (a) History of images and imaging
    (b) Introduction to Computational Imaging
    (c) Course details

Assorted Pixels (Srinivas, 20 mins)   pdf
    (a) Dimensions of imaging
    (b) Multiplexing and Multi-dimensional sampling
    (c) Examples of assorted pixel cameras
        (intensity, color, polarization, depth, focus, angle, time, etc).
    (d) HDR Imaging and Videography.


Coding and Modulation in Cameras (Amit, 45 mins)    pdf
       (a) Photo manipulations versus Coded Imaging
       (b) Key concepts: PSF Invertibility and PSF Invariance
       (c) Techniques for Handling Motion Blur and Defocus Blur
       (d) Strobing Camera: High Speed Imaging for Periodic Signals using Compressive Sensing
       (e) Coded Sampling for Temporal Super-Resolution


Light Fields and Applications (Ankit, 60 mins)  pdf
       (a) Introduction to Light Fields
       (b) Light Field Capture Methods
       (c) Shield Fields
       (d) Glare Reduction

       (e) Re-Interpretable Imaging: Varying Resolution in Post-Capture
       (f) Bi-Di Screen for 3-D Gesture Interaction
       (g) Bokode: Camera based Interaction from a Distance


Computational Illumination (Srinivas, 45 mins)   pdf
        (a) Geometry Acquisition
        (b) Acquiring Material Properties and Appearance
        (c) Acquiring Volumetric Media
        (d) High Speed Vision
        (e) Novel Displays


Future Directions (Amit, 15 mins)  pdf
        (a) Imaging beyond Computer Vision
        (b) Smart Optics, Smart Sensors and Smart Lighting
        (c) New Challenges in Image Authentication, Privacy and Social Impact