Text Detection and Pattern Recognition Based on Deep Learning

In this paper we proposed a fast text detection method in which it is based on deep learning technique of neural network and also clustering method is used in classification or recognition of text. This project is aimed at summarizing and analyzing the major changes and significant progresses of scene text detection and recognition in the deep learning era. The extraction of text from a natural image is a challenging task .Text detection and character recognition in images is a research area which attempts to develop a computer system with the ability to read the text and pattern from images .At first a plain paper containing text is scanned and saved as image (.JPG).This image than pre-processed ,segmented and detection of text done successfully using pixel scanning system . Text can be recognized with and without segmentation of characters. Segmentation can be line, word or character. While without segmentation character is recognized from whole text image. Here we use a new technique to detect text from the image called diagonal based feature extraction in last layer of convolutional neural network with the help of genetic algorithm. Once the extraction of is feature done we provide the training to learning machine. Along with this feature we use feed forward network as a classifier and convolution neural network for feature extractor. We basically use deep learning technique for training and testing. Character recognition convolutional neural network CRConvNet has more layers working of all layer shown in flowchart. One dataset which contain 360 training set data that are all in capital(A-Z) and small(a-z) alphabet, digit(0-9) and some special character are also used.