EVALUATION OF HANDWRITTEN DIGITS USING CNN

Abstract

The objective ot this project is to develop Machine learning algorithm to evaluate handwritten digits ofstudents. This report examines various procedures utilized for the acknowledgment of handwritten digits. Various stages including collection of information, picture handling like the pre-processing stage which includes trimming of images and resizing them and afterward amplification of images and feature extraction will be examined and talked about. At last, the finished result of the calculations will be written in the tool called Jupyter Notebook This venture utilizes the Deep Learning model to foster the product. The fundamental benefit of utilizing this is a procedure is that it gives image classification and identification that is reasonable for digit recognition. Convolution neural network (CNN) is examined and utilized in light ofthe fact that it turns out better for data that are addressed as grid structures. v The framework first returns with the assortment of information and pre-interaction of the gathered pictures with expansion and smoothing. Division, resizing and includes extraction are likewise acted simultaneously. Then, the feed forward process through the organization is summoned to yield a result grid. In view ofthe result network, the perceived person not set in stone. This framework is intended to modify the organization for a singular client. Suggestions for future turn of events and ends additionally remembered for the report.