Abstract:
the
reconstructed images.
This thesis has two thrust: the ¯rst goal is to detect the ring generating stripes
in the sinogram, which encompasses the detection of both isolated and contiguous
stripes generated due to dysfunctional and defunct detector elements. The second
goal is to correct those erroneous detectors' responses in the sinogram domain. As
ring artifacts are manifested as edge creating stripes, single or contiguous, in the
sinogram, they are detected based on the absolute sum of the second derivative of
the sinogram and a self adaptive threshold computed from its ¯rst derivative using
its statistical properties. The wide band contiguous stripes are detected using the
mean curve and multilevel polyphase decomposition of the given sinogram. For
the correction of ring artifacts, variable window moving average (VWMA) and
weighted moving average (WMA) ¯lters are proposed in this work.
The robustness of the proposed method is then checked by producing images
reconstructed from both fan beam and cone beam reconstruction algorithms.
All diverse data and di®erent patterns of artifacts like isolated or band rings,
artifacts from defunct and mis-calibrated detectors are observed in this thesis.
Experimental results show that the proposed method can detect ring artifacts
with high accuracy and thus remove them more e®ectively without imparting
noticeable distortion in the image as compared to other reported techniques.
