OCR Error Rate Versus Rejection Rate for Isolated Handprint Characters

A 1500-CHARACTER OR LESS FACTUAL SUMMARY OF MOST SIGNIFICANT INFORMATION. IF DOCUMENT INCLUDES A SIGNIFICANT BIBUOGRAPHY OR UTERATURE SURVEY. CITE IT HERE. SPELL OUT ACRONYMS ON FIRST REFERENCE.) (CONTINUE ON SEPARATE PAGE, IF NECESSARY.) ~; f}'I.a/yL+'-<-~1.Jl'tA (n~W f...< oQ.....~1 &lt~<-eCi.Lt-~) Over twenty-five~icipating in the First Census OCR SystemsConference submitted confidence data as well as character classification data for the digit test in that Conference. A three parameter function of the rejection rate r was fit to the error rate versus rejection rate data derived from this data, and found to fit it very well over the range from r = 0 to r = 0.15. The probability distribution underlying the model e(r) curve was derived and shown to correspond to an inherently inefficient rejection process. With only a few exceptions that seem to be insignificant, all of th s ste submitting data to the Conference for scoring seem to employ this same rejection process with a remarkable uniformity of efficiency with respect to the maximum efficiency allowed for this process. Human classification of a subset of the digit test suggests that there is considerable room for improvement in the performance of machin~ OCR before the theoretical ideal is achieved. ~ .iA-~J.:.rv} error rate; isolated character; hand print; OCR; segmente~ character. AVAILABIUTY UNUMITED 0 FOR OFFICIAL DISTRIBUTION. DO NOT RELEASE TO NTIS. ORDER FROM SUPERINTENDENT OF DOCUMENTS, U.S. GPO, WASHINGTON, D.C. 20402 ORDER FROM NTIS, SPRINGFIELD. VA 22161 NOTE TO AUTHOR(S) IF YOU DO NOT WISH THIS MANUSCRIPT ANNOUNCED BEFORE PUBUCATION. PLEASE CHECK HERE. 0