Date of Award
In this research, an attempt to create a neural network classifier for the Latin uncial alphabet found in the Codex A, a late fourth century manuscript held by the Archivo Capitolare in Vercelli, Italy and the oldest of the Latin Gospels, is outlined. Training data collection methods, as well as image processing and artificial data creation methods used for expanding the training set, are discussed, as is their contribution to the success of the neural networks. While successful results are achieved for the testing sets, classification of new image data is largely unsuccessful. Potential improvement and avenues for expanding the research are discussed, as are potential explanations for the results.
Anthony, Eleanor C., "A Neural Network Approach to Handwriting Recognition in the Codex A" (2016). Honors Theses. 277.