COMPUTER PROCESSING OF RI IMAGES BY USE OF LOW-PASS FILTERS (II) : FILTERINGS AND SUBTRACTIONS
HIROYUKI HANADA · Japanese Journal of Radiological Technology · 1981
Seven low-pass digital filters which were synthesized previously by the author are applied to 64×64 matrix images of a IAEA liver phantom, which is filled with 9.25×10^7Bq (2.5mCi) ^Tc aqueous solution, from a scintillation camera. Three matrix images are made an object of investigation and total counts of them are 30k counts, 100k counts and 600k counts respectively. First, low-pass digital filterings by convolution are performed to the matrix images and the processed images are observed about image quality. When the filter with the smallest cutoff frequency, which was called Filter A, was used, the fact was shown that almost all the informations in the processed images for three kinds of count data disappeared. Finally, subtractions between the processed image by Filter A and the processed image by another filter are performed for each count data. The consequences of these processings are as follows : (1) The low-pass filtering is effective in improvement of image quality for 30k count image and 100k count image. However, this method does no good in 600k count image. (2) The subtraction results in a success about image restoration with the supression of noise and the reduction of blur, and this method is superior in improvement of image quality to the low-pass filtering. On the other hand the processing conditions for image restoration depend on the count level of the image.