초록

용어보기논문에서 용어와 풀이말을 자동 추출한 결과로, 시범 서비스 중입니다. AI-Helper

Abstract ▼ AI-Helper

Recently, the amount of mails to be delivered by the post office centers increases rapidly. Compared to the internationally active researches in developing automatic mail sorting machines, a little research result has been published in our country. A Korean regular letter envelope contains 6 ZIP co...

Recently, the amount of mails to be delivered by the post office centers increases rapidly. Compared to the internationally active researches in developing automatic mail sorting machines, a little research result has been published in our country. A Korean regular letter envelope contains 6 ZIP code boxes which makes easy the automatic detection and recognition of ZIP code numerals. The target of this paper is to develop a practical system which detects and recognizes the ZIP code numerals on the Korean letter envelopes. Our system consists of three modules, module for extracting the 6 ZIP code numerals, module for recognizing the extracted handwritten numerals, and module for transforming the recognition results into a sorting information. We set up several design criteria before attempting development of the system Algorithms were developed under those design criteria. We performed a performance evaluation with 204 letter envelope images written by University students which include various-size letter envelopes and post cards. The execution time of our system is 0.93 second on a PC which amounts to the sorting speed by 4 persons, The sorting success rates (78% for the first digit, 61% for the first 3 digits, and 41% for all the 6 digits) show that our system can be put to a practical, applications by some tunings to the field environments.