Word Recognition Method Using Convolution Deep Learning Approach Used in Smart Cities for Vehicle Identification

Recognizing words or numbers from an image or a text is a complex issue yet useful in different security applications. Dynamic coordinating between characters of a vocabulary
passage and segment(s) of the information picture is utilized to rank the dictionary sections arranged by best match. In the proposed work words or numbers are recognized from images or text files or handwritten text, as this application can be applied to smart cities for improving security to recognize the new vehicles entering into the city. Traditional strategies frequently neglect to result acceptable outcomes in identification. Hence, in this
proposed work, we propose to consolidate Key Pixel Locator (KPL) in an image and combine it with Convolutional Neural Network (CNN) to accomplish great recognition rate
and identification rate. The characters are identified in a word from the vehicle number plate and the data extracted can be verified to recognize the vehicle and its owner details so that vehicle detection can be easy in case of theft or any criminal vehicle. Exploratory outcomes demonstrate that our methodology utilizing the variable length beats the strategy utilizing fixed span as far as both exactness and speed. Speed of the whole recognition
procedure is around 200 ms and the recognition exactness is 97% is accomplished.