Digit recognition using neocognitron applied to mail zip code identification.

• Main Authors: Padura, Jasper T., Saavedra, Ramil G., Santos, Aldwin A., Sy, Richmond S., Yu, Kenneth L.
• Format: text
• Language: English
• Published: Animo Repository 2000
• Subjects: Optical character recognition devices; Area identification; Zip codes; Electrical and Computer Engineering
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/10265
• Institution: De La Salle University

Mail zip code identification by using the neocognitron neural network to recognize the digit is a step towards improving the mail sorting process. Our present postal system sorts mail manually thereby producing lots of wasted time that could have been converted to useful opportunities. One possible solution is by making the sorting process automated, thus improving the time it takes for the letter to reach its destination. Fast and reliable postal system is what fits our competitive and demanding world today. For our system, one digit of the zip code should be recognized, then sort the mails to a certain pre-programmed arrangement. A video camera will be used to capture the zip code. Then the computer will send signals to the sorting mechanisms, in this case a conveyor and rotating dish both driven by a motor, to direct the mail to its proper box. Interfacing circuits will also be used in order for the computer to know when to begin and end image capturing process. Similarly, a program interface will be created that will facilitate the communications between the computer and the rotating dish. The tricky part of this thesis is the reliability of the algorithm in recognizing and interpreting handwritten and typewritten zip codes.