作者: EIICHI HONDA , MASAHARU DOMON , TAKEHITO SASAKI
DOI: 10.1097/00004424-199110000-00012
关键词: Sierpinski triangle 、 Dilation (morphology) 、 Koch snowflake 、 Personal computer 、 Fractal dimension 、 Mathematics 、 Pattern recognition 、 Pixel 、 Artificial intelligence 、 Fractal dimension on networks 、 Fractal analysis
摘要: A method for measuring the fractal dimension of radiographic images was devised by modified pixel dilation (PD) method, and its accuracy examined two types image analyzers: one controlled a minicomputer based on hexagonal array pixels, other personal computer square pixels. It shown that dimensions measured using analyzers three curves (straight line, Koch curve, Sierpinski gasket) were deviated less than 2.1% from theoretical dimension. The time required measurements fell within 10 seconds minicomputer-controlled analyzer, 12 minutes computer-controlled which contrasts to 30 or more available methods. Analysis ductal patterns in sialograms this indicated significantly different between normal parotids with Sjogren syndrome. PD can be used sialograms. may serve as numerical gradings progression disease.