Optical Recognition of Handwritten Logic Formulas Using Neural Networks
نویسندگان
چکیده
In this paper, we present a handwritten character recognition (HCR) system that aims to recognize first-order logic formulas and create editable text files of the recognized formulas. Dense feedforward neural networks (NNs) are utilized, their performance is examined under various training conditions methods. More specifically, after three algorithms (backpropagation, resilient propagation stochastic gradient descent) had been tested, created trained an NN with descent algorithm, optimized by Adam update rule, which was proved be best, using trainset 16,750 image samples 28 × each testset 7947 samples. The final accuracy achieved 90.13%. general methodology followed consists two stages: processing design training. Finally, application has implements automatically recognizes An interesting feature it allows for creating new, user-oriented sets parameter settings, thus new models.
منابع مشابه
Handwritten numeral recognition using autoassociative neural networks
nn ee tt ww oo rr kk s ee ss ss ,, pp ll ee aa ss ee .
متن کاملHandwritten Character Recognition Using Neural Networks
The aim of this work is to assess the possibility of having on-line handwriting recognition abilities on a small portable computer. We studied the recognitions rates dependence on the target character set and the incremental training made possible by new neural network techniques. We worked on cross-cultural data gathered from French and Dutch people containing about 70000 characters from 200 p...
متن کاملPersian Handwritten Digit Recognition Using Particle Swarm Probabilistic Neural Network
Handwritten digit recognition can be categorized as a classification problem. Probabilistic Neural Network (PNN) is one of the most effective and useful classifiers, which works based on Bayesian rule. In this paper, in order to recognize Persian (Farsi) handwritten digit recognition, a combination of intelligent clustering method and PNN has been utilized. Hoda database, which includes 80000 P...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Electronics
سال: 2021
ISSN: ['2079-9292']
DOI: https://doi.org/10.3390/electronics10222761