A Real-Time Character Recognition System
Ben Shelton · 2012
The paper discusses the implementation of a character recognition system that can be used to effectively translate images taken from a webcam to a digital document in real time. Artificial neural networks provide a great platform for character recognition that allows for quick learning and recognition rates to provide real time functionality. The program discussed uses resilient propagation a feed-forward, supervised learning, neural network algorithm that allows for quick training on any size data set for high functionality. With initial tests producing results of up to 92% correct recognition on a range of varied length training sets.