In the quest to study human perception, the field of character recognition remains prominent. Research has progressed from early, cumbersome optical character readers used for the recognition of a limited set of individual characters to the present-day automatic recognition of handwritten cursive script. This paper attempts to clarify the fundamentals of character recognition, highlighting the processes involved in using a standard database for 'learning' character sets and also the standards and constraints imposed by researchers concerning the constitution of a valid character. A number of various feature extraction techniques which enable individual characters to be recognized are discussed and compared. © 1986.
|Number of pages||11|
|Journal||Image and Vision Computing|
|Publication status||Published - Nov 1986|
- feature extraction
- handwritten characters
- pattern recognition