Objective - The final objective of the project is to recognise the images of the capital handwritten characters and store it in a file to display.
Tools used - Image processing tool box and the neural network training Tool box of MATLAB.
How to run -
Step1 > Load the trained.mat ( It is the dataset in which characteristic of all the capital characters images are stored. It is loaded characteristics of 2000 images. Charcteristics are exracted on the basis of dct2 ( Discrete Cosine Transfor ). Dataset not uploaded.)
Step2 > Run the extractor.m
Step3 > Choose the image using the Gui created.
Step4 > Run the final.m file.
Step5 > Open the file result.txt for output. (Output can varry from original because 2000 images dataset is not enough to recognize all the character images.)
OR
Step1 > Run the main_program.m
Step2 > Choose the image using the Gui created.
Step3 > Run the final.m file.
Step4 > Open the file result.txt for output. (Output can varry from original because 2000 images dataset is not enough to recognize all the character images.)
Working -
File1 > extractor.m ( select the image ) line no 88 to 112 useful code and defined working with comments after each command.
File2 > rendering.m ( filter the image ) line no 6 to 75 useful code and defined working with comments after each command.
File3 > character_extracter.m ( extract the character from the filtered image and get its dct charcteristics in an array(final_array) and store it in base work space.) line no 5 to 45 useful code and defined working with comments after each command.
File4 > final.m ( check the retrived characteristics in dataset and store the result in result.txt ) line no 8 to 186 useful code and defined working with comments after each command.
NOTE : nntrainer.m is to create and train the neural network with extracted features (characteristics) of input images and output required.