Off-line Handwriting Recognition from Forms

A public domain optical character recognition (OCR) system has been developed by the National Institute of Standards and Technology (NIST) to provide a baseline of performance on off-line handwriting recognition from forms. The system’s source code, training data, and performance assessment tools are all publicly available. The system recognizes the handprint written on Handwriting Sample Forms as distributed on the CD-ROM, NIST Special Database 19. The public domain package contains a number of significant contributions to OCR technology, including an optimized Probabilistic Neural Network (PNN) classifier that operates a factor of 20 times faster than traditional software implementations of this algorithm. The modular design of the software makes it useful for training and testing set validation, multiple system voting schemes, and component evaluation and comparison. As an example, the OCR results from two versions of the recognition system (each using a different method of form removal) are presented and analyzed. It is shown that intelligent form removal can improve lowercase recognition by as much as 3%, but this net increase in performance is insufficient to understand the impact on the recognition. A method of analysis is provided, whereby the local changes in performance (both gains and losses) of the system are automatically determined. This involves analyzing the distribution statistics of corresponding character confusion pairs between the two systems. As off-line handwriting recognition technology continues to improve, sophisticated analyses like this are necessary to reduce the errors remaining in complex recognition problems.