Artificial intelligence optical character recognition system and method

A kind of artificial intelligence(AI) optical character recognition (OCR) system and method. The way it operates mainly includes: intercepting each text image range first from the texts selected, preparing two (or more) identical neural network architecture units, with the same weight parameters of each neural network, inputting similar or different text images respectively, and comparing the calculation results after calculations to see if the weights are similar. If the similarity was as high as the set standard value, they would be classified as the same type of text, otherwise different. Through such procedures, the training samples were gradually divided into text sets of different textual meanings, becoming a complete AI OCR system. For text detection needs, the system can increase the training sample data by comparing the texts, without increasing the training set. Simultaneously, it can improve the flexibility of recognizing test texts. The same set of training models can correspond to different fonts