Title :
Three-dimensional non-linear image filter for improving CT images
Abstract :
We have developed a 3D non-linear image filter to provide higher image resolution while maintaining same level of overall image noise for various high resolution CT applications such as IAC, spine, cervical and stent assessments. The filter performs continuous segmentation of the images based on the CT numbers. A gain curve is generated which is a function of image CT numbers that can be negative for certain CT numbers. A deconvolution operation was performed on the images and was then modulated using the gain curve to effectively obtain different degrees of smoothing and resolution enhancement to boost image resolution for high-density objects while maintaining or reducing overall image noises.
Keywords :
bone; computerised tomography; deconvolution; image enhancement; image resolution; image segmentation; medical image processing; nonlinear filters; CT images; IAC assessment; cervical assessment; continuous image segmentation; deconvolution; high image resolution; high-density objects; overall image noise; resolution enhancement; smoothing methods; spine assessment; stent assessment; three-dimensional nonlinear image filter; Computed tomography; Deconvolution; Filters; Image resolution; Image segmentation; Imaging phantoms; Noise level; Noise reduction; Performance gain; Smoothing methods;
Conference_Titel :
Nuclear Science Symposium Conference Record, 2003 IEEE
Print_ISBN :
0-7803-8257-9
DOI :
10.1109/NSSMIC.2003.1352580