Method of and system for detecting anomalies in projection images generated by computed tomography scanners
View Patent ↗A method of and a system for detecting anomalies in projection images generated by CT scanners are provided. One type of anomaly of particular interest is bright or/and dark dots in projection images, which correspond to streak artifacts in the CT images. The method for detecting such bright or/and dark dots in projection images comprises: generating projection images; computing a CFAR distance map; computing a preliminary dot map; generating dot histograms; and detecting bright dots or/and dark dots based on the generated histograms.
1. A method of detecting bad detector channels of a detector array mounted on a rotating gantry of a computed tomography (CT) scanner comprising:
A. Generating two-dimensional projection images using said CT scanner, wherein said two-dimensional projection images are generated during a normal operating mode of said CT scanner without using an image quality phantom;
wherein said two dimensions include the transportation direction of a scanned object (Z-axis);
B. Calculating a two-dimensional constant false alarm rate (CFAR) distance map using said two-dimensional projection images generated from said CT scanner;
C. Generating a two-dimensional ternary-valued preliminary dot map from said CFAR distance map by comparing said CFAR distance map with a pre-defined threshold and with the negative of said pre-defined threshold;
D. Generating a two-dimensional dot histogram using said ternary-valued preliminary dot map: wherein the size of one of said two dimensions is the transportation distance along the Z-axis corresponding to one gantry rotation; and
E. Detecting bad detector channels by comparing said two-dimensional dot histogram with a pre-defined threshold and by using the locations of those dots beyond said pre-defined threshold.
2. A system for detecting bad detector channels of a detector array mounted on a rotating gantry of a computed tomography (CT) scanner comprising:
A. A component constructed and arranged so as to generate projection images using said CT scanner, wherein said two-dimensional projection images are generated during a normal operating mode of said CT scanner without using an image quality phantom; wherein said two dimensions include the transportation direction of a scanned object (Z-axis);
B. A component constructed and arranged so as to calculate a two-dimensional constant false alarm rate (CFAR) distance map using said two-dimensional projection images generated from said CT scanner;
C. A component constructed and arranged so as to generate a two-dimensional ternary-valued preliminary dot map from said CFAR distance map by comparing said CFAR distance map with a pre-defined threshold and with the negative of said pre-defined threshold;
D. A component constructed and arranged so as to generate a two-dimensional dot histogram using said ternary-valued preliminary dot map: wherein the size of one of said two dimensions is the transportation distance along the Z-axis corresponding to one gantry rotation; and
E. A component constructed and arranged so as to detect bad detector channels by comparing said two-dimensional dot histogram with a pre-defined threshold and by using the locations of those dots beyond said pre-defined threshold.