Qiao Hu

Focus detection from digital in-line holograms based on spectral l1 norms.

A rapid focus-detection technique based directly on the spectral content of digital holograms is developed. It differs from previous approaches in that it does not need a full reconstruction of the image. The technique uses l(1) norms of object spectral components associated with the real and imaginary parts of the reconstruction kernel. Further, the l(1) norms can be computed efficiently in the spatial frequency domain using a polar coordinate system, yielding a drastic speedup of approximately 2 orders of magnitude compared with image-based focus detection. Significant computational savings are achieved when subsequent image reconstructions are done selectively over the detected focus distances. Focus-detection results from holograms of plankton are demonstrated that show the technique is both accurate and robust.

Advances in Plankton Imaging Using Digital Holography

An in situ plankton imaging system using a high space-bandwidth product camera and fiber coupled diode laser for digital holography was developed. High quality imaging results and solutions for associated computational challenges are discussed.

Rapid extraction of 3D regions of interest from digital holograms

Selective reconstruction with data reduction is the key to efficient data analysis and high quality image reconstructions from digital holograms. This paper presents an efficient method of extracting 3D regions of interest (ROIs) of object volume directly from a hologram. The new method consists of a rapid focus detection algorithm, a summation kernel for computing a sum image onto which of a sequence of object images are projected and an image segmentation algorithm. The focus detection algorithm is based on the h norm of the object spectral components and works directly on the hologram. It estimates the depth distances at which objects can be found in-focus. The summation kernel is then employed to obtain sum images that are effectively the parallel projection of object images located within each detected range of focusing distances. The lateral support of these ROIs are then obtained by segmenting these sum images. The whole process does not require full reconstruction of the object images throughout the 3D volume. Based on a lens-less DHI camera system jointly developed at MIT and WHOI, the new 3D ROI extraction technique has been applied to plankton holograms obtained in the laboratory and on shipboard. Results from plankton holograms are demonstrated.

Fast Computation of Focal Planes for Sparsely Populated Digital Holograms Using a Spectral l 1 Norm

A new focus measure utilizing l1norms in the Fourier domain provides focal depth estimates without reconstructing an object image. A polar implementation reduces the problem to one dimension, yielding two orders of computational speedup.