Deep Learning Model for character recognition.
This repository contains all the codes and reference data for building the Handwritten Character recognition Model from scratch.
A Deep Learning Model for handwritten character recognition (A-Z). The Dataset containg 26 folders from A to Z containing handwritten images in size 28*28 pixels, each alphabet in the image is centre fitted. Each Image is stored as Gray-level. The feature extraction technique is obtained by normalizing the pixel values. Pixel values will range from 0 to 255 which represents the intensity of each pixel in the image and they are normalized to represent value between 0 and 1. Convolutional neural network is used as a classifier which trains the kaggle dataset. The prediction for the given input image is obtained from the trained classifier.
data.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 372450 entries, 0 to 372449
Columns: 785 entries, 0 to 0.648
dtypes: int64(785)
memory usage: 2.2 GB
data.shape()
(372450, 785)
- (A-Z) Images : Contains the Handwritten images written by me;
- best_model.h5 : Deep Learning Model;
- Character recoginition.ipynb : Code for building the model from scratch;
- Hand Written character Recognition.ipynb : contains the code for recogizing the handwritten characters written by me.
- Real Time Character Recognition
- Better Image Processing such as: reduce background noise to handle real time image or images in different lighting conditions more accurately.
Feel Free to suggest an improvemet in this project.
This is a work from my 4th semester in Computer Enginnering from Saffrony Institute of Technology. Also I was assigned this project work from the Nano Degree course of Autonomous Vehicles by Elite Techno Groups in April 2022.