HAND JOTTED NUMERAL REALISATION USING CNN

HAND JOTTED NUMERAL REALISATION USING CNN

Ms. Rajashree Sutrawe, D.Shyam Sundar,G.Kuladeep Nayyar

CSE, Guru Nanak Institutions Technical Campus, Hyderabad, Telangana, India

ABSTRACT

The issue of verifying registration numbers has remained an open question in legal standards for years. Some believe neural systems possess remarkable capabilities in manipulating information. This article aims to propose an effective strategy for confirming the emergence of composite content by analyzing available models. It focuses on Convolutional Neural Networks (CNNs), which are known for their efficiency but can sometimes weaken the framework. In machine learning, neural systems can be utilized for tasks such as code traversal. By utilizing the MNIST database and integrating it with CNNs, a clear progression strategy was developed. Essential libraries like NumPy, Pandas, TensorFlow, and Keras are key to this modeling approach. The MNIST dataset contains about 70,000 images of digits 0-9, categorized into 10 classes. This data is split into training and test sets, with each image represented as a 28x28 grid of grayscale pixels.

Keywords: Data processing,Convolutional neural networks(CNN),MNIST(Modified National Institute of Standard and Technology),TensorFlow,Keras,Numpy,Pandas