Abstract:
The use of °at-panel detectors (FPDs) is becoming increasingly popular in cone
beam volume and multi-slice CT imaging. But due to the presence of the defective
and insu±cient calibrated detector elements in the FPD, the diagnostic quality of
the FPD-based CT images in both CT systems is degraded by ring and radiant
artifacts. In this thesis, a detail analysis including the classi¯cation, detection and
correction of these ring artifacts is presented. Unlike conventional approaches, we
emphasize here on the separate detection and correction schemes for each type of
rings for their e®ective removal. Two di®erent ring and radiant artifact correction
algorithms are presented here: one is applicable for the fan or parallel beam geometry
based multi-slice CT and the other for the cone beam geometry based CT. For
the correction of the ring artifacts in multi-slice CT, an artifact strength based
sinogram-processing method is presented and a mathematical index is used to isolate
the defective and mis-calibrated pixels from the good ones. For the correction of
artifacts resulting from the defective detector elements, 2D variable window moving
average and weighted moving average ¯lters are proposed. On the other hand, the
normalization correction scheme is adopted to correct the responses of the mis-
calibrated detector elements. Next, a novel approach is presented to eliminate the
ring and radiant artifacts of a 3-D cone beam volume CT image using the proposed
sinogram based ring artifact removal method.
Finally, an e®ective and fast ring artifact correction technique speci¯cally for 3-D
cone beam volume CT image is proposed. Some template images are derived from
the responses of the detector pixels using their statistical properties and then, an
e®ective non-causal derivative based detection algorithm in 2-D space is presented for
the detection of defective and mis- calibrated detector elements separately. An image
inpainting based 3-D correction scheme is proposed for the estimation of responses of
defective pixels and the responses of the mis-calibrated pixels are corrected using the
normalization technique. A simpli¯cation of the proposed method is also suggested
for its real time implementation. The overall experimental results demonstrate the
superiority of the proposed two correction methods (applicable for the multi-slice
and cone beam CTs) over the other methods reported in the literature.
