OCR on binary image

Hello, I have an image and I've pre-processed it to be as clear as possible to get a binary image - for OCR.

However, the OCR function does not recognize any character - how can it be improved for my case?

This is what I am doing:

img_for_ocr = imread("img.png");
ocrResults = ocr(img_for_ocr(:,:,1));
figure, imshow(img_for_ocr(:,:,1))
fprintf("Recognized text: %s \t Detection confidence: %0.4f \n",cell2mat(ocrResults.Words),ocrResults.WordConfidences)

Answer 1

[Image Analyst]

Here's an improvement, though I don't have time to finish it now.

% Demo by Image Analyst
clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
clear;  % Erase all existing variables. Or clearvars if you want.
workspace;  % Make sure the workspace panel is showing.
format long g;
format compact;
fontSize = 16;
markerSize = 20;
%--------------------------------------------------------------------------------------------------------
%    READ IN TEST IMAGE
folder = [];
baseFileName = 'ocr1.png';
fullFileName = fullfile(folder, baseFileName);
% Check if file exists.
if ~exist(fullFileName, 'file')
    % The file doesn't exist -- didn't find it there in that folder.
    % Check the entire search path (other folders) for the file by stripping off the folder.
    fullFileNameOnSearchPath = baseFileName; % No path this time.
    if ~exist(fullFileNameOnSearchPath, 'file')
        % Still didn't find it.  Alert user.
        errorMessage = sprintf('Error: %s does not exist in the search path folders.', fullFileName);
        uiwait(warndlg(errorMessage));
        return;
    end
end
grayImage = imread(fullFileName);
% Get the dimensions of the image.
% numberOfColorChannels should be = 1 for a gray scale image, and 3 for an RGB color image.
[rows, columns, numberOfColorChannels] = size(grayImage)
if numberOfColorChannels > 1
    % It's not really gray scale like we expected - it's color.
    fprintf('It is not really gray scale like we expected - it is color\n');
    % Extract the blue channel.
    grayImage = grayImage(:, :, 3);
end
%--------------------------------------------------------------------------------------------------------
% Display the image.
subplot(2, 2, 1);
imshow(grayImage, []);
impixelinfo;
axis('on', 'image');
title('Original Gray Scale Image', 'FontSize', fontSize, 'Interpreter', 'None');
% Update the dimensions of the image.
% numberOfColorChannels should be = 1 for a gray scale image, and 3 for an RGB color image.
[rows, columns, numberOfColorChannels] = size(grayImage)
% Maximize window.
g = gcf;
g.WindowState = 'maximized';
g.Name = 'Demo by Image Analyst';
g.NumberTitle = 'off';
drawnow;
%--------------------------------------------------------------------------------------------------------
% Get mask by thresholding at 128.
lowThreshold = 128;
highThreshold = 255;
% Interactively and visually set a threshold on a gray scale image.
% https://www.mathworks.com/matlabcentral/fileexchange/29372-thresholding-an-image?s_tid=srchtitle
% [lowThreshold, highThreshold] = threshold(lowThreshold, highThreshold, grayImage)
mask = grayImage >= lowThreshold & grayImage <= highThreshold;
caption = sprintf('Histogram of Image.  Threshold at %d', lowThreshold);
title(caption, 'FontSize', fontSize, 'Interpreter', 'None');
%--------------------------------------------------------------------------------------------------------
% Display initial mask.
subplot(2, 2, 2);
imshow(mask);
impixelinfo;
axis('on', 'image');
title('Initial Mask Image', 'FontSize', fontSize, 'Interpreter', 'None');
drawnow;
%--------------------------------------------------------------------------------------------------------
% Clean up the initial mask -- get rid of small interior holes in the image.
holeMask = imclearborder(~mask);
%--------------------------------------------------------------------------------------------------------
% Optional: find areas of blobs so we can filter out small ones.
props = regionprops(holeMask, 'Area');
allAreas = sort([props.Area])
% Get mask of just the small holes.
holeMask = bwareafilt(holeMask, [1, 600]);
%--------------------------------------------------------------------------------------------------------
% Display hole mask.
subplot(2, 2, 3);
imshow(holeMask);
impixelinfo;
axis('on', 'image');
title('Hole Mask Image', 'FontSize', fontSize, 'Interpreter', 'None');
drawnow;
% Fill in those holes in the original mask
mask(holeMask) = true;
% Get rid of blobs smaller than a letter.
% props = regionprops(mask, 'Area');
% allAreas = sort([props.Area])
mask = bwareaopen(mask, 1000);
%--------------------------------------------------------------------------------------------------------
% Display the final mask
subplot(2, 2, 4);
imshow(mask);
impixelinfo;
axis('on', 'image');
title('Final, Filled Mask Image', 'FontSize', fontSize, 'Interpreter', 'None');
txt = ocr(mask)
fprintf('The text is %s.\n', txt.Words{end})

[Saania]

Thank you for your response. I see that the snippet you shared basically tries to cover up the holes, however, I don't see this as the only issue. Covering up the holes also covers up characters like "B" or "8". In the image above, C, B and 3 are clear - OCR should detect atleast these three characters. The image below: does not have any holes - OCR still doesn't work (even for CB1-2).

[Image Analyst]

Yeah, we know the OCR algorithm they use is not the best. A few years ago they put it on the list of functions that needed another looking at but I guess tehy haven't gotten around to it yet. Maybe you can try deep learning. Look for their MNIST demo where they're using a deep learning model to predict numbers and (possibly) alphabetic characters.

[Saania]

This brings a closure, I was wondering if I am not calling the function and it's parameters appropritely. Thankyou!