Methods: Analyzing the curvature algorithm for assessment of vascular tortuosity was based on the real and schematic images. Since the algorithm had an appropriate behavior toward the schematic images, it was applied to real ones. Curvature algorithm estimation in this study had a close result to previous studies with the privilege of being less time consuming and less sophisticated. To measure vascular diameter, we designed an algorithm !"#$%&'%$(")!'*#%)+!'",&+-.%/'%)0("%!12&+()0-3%4+")153% !"#$%&'%)0#%6!"*71!)7+#%(-!2#%)+!'",&+-%)0#%$(")!'*#%8!"%(-91(*!)#$%!'$%"#*&'$153%8()0%!'!- 15:('2%)0#%+#"71)"%)0#%$(")!'*#%8!"%-#!"7+#$.%;&%!""#""%)0("%!7)&-!)#$%!12&+()0-3%<=%(-!2#"%)!>#'%8()0%?#)*!-%8#+#%*1!""(4#$%)&%)8&%2+&79"%&,%917"% and non plus patients by 3 experts. Results:%/'%)0#%4+")%91!*#3%)0#%!**7+!*5%&,%!12&+()0-%8!"%$(6($#$%(')&%)0+##%2+&79"%!"%,&11&8"@% #('2%*&-9!)( 1#%8()0%!)%1#!")%&'#%&,%)0#%#A9#+)"3% being compatible with at least two of them, and being compatible with all of the experts diagnosis. In this part, it was noted that in 87.5 % of cases the algorithm was compatible with at least one of the expert diagnosis in plus or non plus patients. Because in statistical analysis the expert (3 ) had highest discriminatory power in segregation of Plus and Non-Plus Patients , on second evaluation this automated algorithm results was compared with expert ( 3 ). The threshold values for the two parameters, tortuosity and dilatation were chosen in such a way that maximum accuracy in detecting a fundus (-!2#%!"%917"%$("#!"#% #%!*0(#6#$%.%B,)#+%!'!15:('2%)0#%!6#+!2#%6!17#"%3%"#*&'$%!""#""-#')%0!$%"#'"()(6()53%"9#*(4*()5%!'$%!**7+!*5%&,%C.=<DE3%C.CFGEH% and 0.7726, respectively in comparison of expert 3 . Conclusion: this New Automated Algorithm has acceptable Accuracy for Plus disease Detection.

Keywords