Infotech Oulu Graduate School

Digital hologram image processing

---

Duration: 20 hours, with additional optional laboratory work

Given by Thomas Naughton (Senior Lecturer, National University of Ireland, Maynooth, and Marie Curie Fellow, University of Oulu) and Dr. Bryan Hennelly (IRCSET Fellow, National University of Ireland, Maynooth)

Abstract for the course

Digital holography captures the full wavefront (amplitude and phase) of the laser light reflected from a real-world three-dimensional (3D) scene using a digital camera without imaging optics. Real-valued intensity images corresponding to different focal lengths, different depths of field, and even different perspectives of the 3D scene can be calculated numerically from a single digital hologram by a virtual lens arrangement. In this sense, we regard a digital hologram as being a generalisation of a digital image. Digital holographic image processing is a new field of research that tries to define and invent versions of tools from traditional image processing for this new digital medium. In this course we will introduce digital holography to a computer science/electrical engineering audience. The course will begin with a brief introduction to the associated hardware (lasers, LCD panels, and CCD sensors), the equations governing the evolution of a coherent (laser-generated) image as it propagates in free space (the Fresnel and other integrals), how conventional holograms are obtained, and why holography, in principle, is regarded as the definitive non-contact method of capturing 3D information about an object. This will be followed by various arrangements for capturing a hologram with a digital camera (a digital hologram), formulations of discrete versions of the propagation equations, and demonstrations of how a digital hologram can have perspectives of its encoded 3D object reconstructed in software. The second part of the course will concern the new field of holographic image processing: the application of image processing to digital holograms. This field differs to conventional image processing in three main respects: (1) digital holograms are complex-valued images, (2) they contain a multiplicative noise-like effect called speckle, and (3) in holographic image processing we work in the Fresnel domain rather than the space domain. All existing aspects of holographic image processing will be covered, including twin-image removal, holographic microscopy and pure phase objects, hologram compression and encryption, speckle reduction, wavelet approaches, superresolution techniques, capturing holographic video and the holographic video standard, segmentation of 3D objects, extraction of 3D surface information, and optical and digital holographic displays. The course will finish on the prospects of digital holography in industrial (e.g. metrology) and consumer applications (e.g. 3D television), and a comparison with competing 3D sensing and display technologies (such as integral imaging, confocal microscopy). The first part of the course will be drawn from textbooks, and second part of the course will be drawn from research papers, all published within the past 5 years. The pre-requisites for this course are basic image processing concepts such as 2D Fourier transform and convolution.

Dates: between 20.08.2007 and 31.08.2007

Location: University of Oulu main campus, August 20-21 in room TS126

Times:

The first six lectures will take place on Monday 20.08.2007 and Tuesday 21.08.2007 from 09:00 to 12:00 each day, in room TS126. The remaining 14 lectures will take place within a two week period, but the exact times remain flexible. The times will be decided during the first lecture to best suit those attending the course.

In order to be sure we bring enough photocopies of the notes, it would be helpful if those intending to attend would send an email to thomas.naughton(at)oulu.fi.

---

Breakdown of the course (by hour):

1 Basic coherent light (coherence, interference, lasers)
2 Scalar diffraction theory and Fourier optics
3 same as above
4 Conventional holography
5 same as above
6 Digital recording - sampling theory
7 Examples of digital hologram pre-processing
8 Reconstructions of perspectives of 3D objects from digital holograms
9 same as above
10 Overview of applications of digital holography
11 Optical display (including computer-generated holograms)
12 Speckle, optical & signal processing speckle reduction techniques
13 same as above
14 Accessing 3D object information encoded in digital holograms
15 Digital holographic microscopy of dynamic objects
16 Watermarking of 3D objects encoded in digital holograms
17 Encryption of 3D objects encoded in digital holograms
18 Metrology applications of digital holography
19 3D displays
20 Competing 3D approaches (confocal microscopy, integral imaging)

It will be useful for participants to complete the optional laboratory work that follows demonstrations during lectures. Software (written for Matlab and requiring the Image Processing toolbox) and example digital holograms will be provided. Although not essential, if students can bring laptop computers (with Matlab + Image Processing toolbox) with them to lectures they will get maximum from the course.

It will be possible to devise self-contained hologram image processing research projects for students who require such research projects for another course. These projects can range from short reports to scientific work leading to OSA/IEEE journal submission. Limited supervision can be provided (T. Naughton will be at the Department of Information Processing Science, University of Oulu, for at least 12 months).

---