Optical character recognition using artificial neural networks
Institution: | Colorado State University – Pueblo |
---|---|
Department: | |
Year: | 2016 |
Keywords: | Optical character recognition; Python (Computer program language); Neural networks (Computer science) |
Posted: | 02/05/2017 |
Record ID: | 2079112 |
Full text PDF: | http://hdl.handle.net/10217/172439 |
Optical character recognition is a complicated task that requires heavy image processing followed by algorithms used to convert that data into a recognized character. While programs exist that already can perform character recognition, they require intensive processing that is not always necessary because they recognize a wide range of characters spanning numerous fonts. In applications where a specific character set in a specific font is defined, the processing requirements can be easily reduced by developing a OCR system tailored to those specifications. One way to do this is with artificial neural networks. This method has many tunable parameters, and in many cases the optimal settings may need to be determined through trial and error. This purpose of this project was to develop and train an artificial neural network on the Raspberry Pi microcontroller and determine the optimal layer settings to identify a specific set of characters. An additional goal was for the coding to be customizable and easy to use so it could be used for other recognition tasks as well beyond just character recognition. Advisors/Committee Members: DePalma, Jude L. (advisor).