High-sensitivity dynamic coded aperture imaging

Coded aperture imaging aims at improving the sensitivity of standard imaging optics (pinholes and parallel-hole collimators) by using hundreds of pinholes. Since each pinhole projects a copy of the object onto the detector, copies overlap. The overlap can be undone by computer post-processing with a simple correlation if pinholes are arranged in particular patterns. These are designed so that, when the object is infinitely far from the detector, a perfect image is obtained. In applications such as nuclear medicine and small-animal imaging the object is kept close to the detector to maximize sensitivity and use of the same patterns was shown to result in noticeable artifacts. The existing technique to mitigate near-field artifacts requires taking two images of the object sequentially with two similar masks, but this makes the implicit assumption of a static object. In near-field imaging, however, it is possible to place two detectors on opposite sides of the object and acquire the two images simultaneously. This approach extends the technique to dynamic studies. It also has the benefit of doubling the overall count rate by better utilization of the heads of a clinical gamma camera. Experimental results show that coded aperture optics can follow the movement of a 370-kBq point source at 0.05 s per frame with 3.5-mm resolution over a 12/spl times/12 cm/sup 2/ field of view. Images of a disk source demonstrate that, as in static studies, near-field artifacts can be almost eliminated.