The FDK algorithm is the most popular cone beam algorithm in the medical and industrial imaging field. Due to data insufficiency acquired from a circular trajectory, the images reconstructed by the FDK algorithm suffe...The FDK algorithm is the most popular cone beam algorithm in the medical and industrial imaging field. Due to data insufficiency acquired from a circular trajectory, the images reconstructed by the FDK algorithm suffer from the intensity droping with increasing cone angle. To overcome the drawback,a modified FDK algorithm is presented by convert the 1D ramp filtering direction from along the horizontal lines to along the nutating lines based on the result of Turbell. Unlike Turbell's method,there is no need for our algorithm to rebin the cone-beam data into 3D parallel-beam data before reconstructing. Moreover pre-weighting of the projection data is corrected by compensating for the cone angle effect. In addition,another correction term derived from the result of Hu is also induced into our algorithm. The simulation experiments demonstrate that the final algorithm can SUDDress the intensity droD associated with the FDK algorithm.展开更多
基金This work was sponsored by Youth Foundation Project of Liaon-ing Province Education Department .(NO.2004F057)
文摘The FDK algorithm is the most popular cone beam algorithm in the medical and industrial imaging field. Due to data insufficiency acquired from a circular trajectory, the images reconstructed by the FDK algorithm suffer from the intensity droping with increasing cone angle. To overcome the drawback,a modified FDK algorithm is presented by convert the 1D ramp filtering direction from along the horizontal lines to along the nutating lines based on the result of Turbell. Unlike Turbell's method,there is no need for our algorithm to rebin the cone-beam data into 3D parallel-beam data before reconstructing. Moreover pre-weighting of the projection data is corrected by compensating for the cone angle effect. In addition,another correction term derived from the result of Hu is also induced into our algorithm. The simulation experiments demonstrate that the final algorithm can SUDDress the intensity droD associated with the FDK algorithm.