SURVEY ON LETTER RECOGNITION USING DEEP LEARNING









Abstract

Handwritten character recognition is famous because of its wide scope of utilization. Processing application structures, digitizing antiquated articles, postal location processing, bank check processing and numerous others are the developing fields in territory of handwritten character processing. Handwritten character is drawing in researchers since last 3 decades. Numerous methodologies have been proposed for successful recognition. The programmed recognition of handwriting has been under scrutiny since the 1950's when applications for the moderately new digital computer technology happened to revenue. From that point forward, there has been consistent research exertion into the programmed processing and recognition of handwriting. The overall classes into which the processing of handwriting can be isolated are introduced. Empowering progress has been accounted for in the programmed recognition of numerical formulae, printed characters, cursive content and mark confirmation, and there has been a restricted measure of research into applying programmed picture processing to individual recognizable proof and record validation through essayist ID or the location of masked or fashioned handwriting. Notwithstanding the recognition of Roman scripts and symbols there has been escalated work into the recognition of Chinese, Japanese, and other handwritten scripts. Different utilizations of handwriting recognition are workable for altering, commenting on and other profoundly intuitive exercises which require pointing and explanation. In this day and age it has gotten simpler to prepare profound neural organizations on account of accessibility of enormous measure of information and different Algorithmic developments which are occurring. Now a-days the measure of computational force expected to prepare a neural organization has expanded because of the accessibility of GPU's and other cloud based administrations like Google Cloud stage and Amazon Web Services which give assets to prepare a Neural Network on the cloud. . Handwriting recognition has picked up a great deal of consideration in the field of example recognition and AI because of its application in different fields. Optical Character Recognition (OCR) and Handwritten Character Recognition (HCR) has specified domain to apply. Different strategies have been proposed to for character recognition in handwriting recognition framework. Numerous examinations and papers portrays the strategies for changing over literary substance from a paper report into machine meaningful structure. In coming days, character recognition framework may fill in as a vital factor to establish a paperless climate by digitizing and processing existing paper records. This paper presents a definite audit in the field of Handwritten Character Recognition. Keywords: Handwritten Character Recognition, Optical Character Recognition.


Modules


Algorithms


Software And Hardware

• Hardware: Processor: i3 ,i5 RAM: 4GB Hard disk: 16 GB • Software: operating System : Windws2000/XP/7/8/10 Anaconda,jupyter,spyder,flask Frontend :-python Backend:- MYSQL