Text recognition method and device and electronic equipment

• Main Authors: LIU ZHENGZHEN, HUANG WEI
• Format: Patent
• Language: Chinese; English
• Published: 22.09.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention discloses a text recognition method, belongs to the technical field of optical character recognition, and helps to improve the text recognition accuracy. The method comprises the steps that a target image is input into a convolutional neural network in a pre-trained character recognition model, a feature map with the height being D and the width being n output by the convolutional neural network for the target image is acquired, and D and n are integers larger than 1; recombining the feature map to obtain a feature sequence of the target image; performing coding mapping on the feature sequence through a sequence coding network in a character recognition model to obtain a coding sequence; and decoding the coding sequence through a CTC decoder in the character recognition model to obtain a character recognition result of the target image. According to the method, the feature map with the height larger than 1 is extracted, so that character recognition can be carried out based on features with fine