Hand Written Text Recognition using Tensor flow

The proposed system presents an innovative method for handwritten character detection using deep neural networks. It is an image segmentation based Handwritten character recognition system. This system uses an OpenCV for performing image processing using Tensor Flow for training the Neural Network. This Neural Network model recognizes the text contained in the image of segmented words. As this words-image is smaller than the image of complete text-lines, the Neural Network can be kept small and training on the CPU is feasible. 3/4 of the words from the validation-set are correctly recognized and the character error rate is around 10%. This system is capable to recognize the text written by the hand of a person. The system is divided into two parts. The first one is the android module, which is used for generating the input for the system. The other one is a server, which will perform the main operation of detecting the character of the text. In the System, the android device takes the input in the image form and send it to the server. The server takes input and starts evaluation, classification and prediction for the image data to find the best match word of the text document. The server creates the text document at last as the output and sends it back to the device. Then the android device shows the output received from the server.