I want extract all the text from the given image.

I am trying to extract the numbers from this image along with its name (FiO2, Pmean, PEEP, etc.). I am using the below code to extract it using the color thresholder app. I am facing 2 problems. Using the color thresholder app i am only able to isolate the blue text (numbers) so the varibale names are getting erased. Secondly, among the blue numbers, I am getting an irrelevant text (GRY) instead of 5.0. What is going wrong?

img: original image

BW: exported image from color thresholder

% Define the path to your image

imagePath = 'thumbnail_6.jpg';

% Read the image

img = imread(imagePath);

% Display the original image to ensure it's loaded

imshow(img);

imshow(BW);

% Perform OCR on the binary image

results = ocr(BW);

% Set text properties

text = results.Text; % Text to be inserted

position = [500, 100]; % Position of the text

fontColor = 'black'; % Color of the font

fontSize = 34; % Font size

% Insert the text into the image

imgWithText = insertText(img, position, text, 'TextColor', fontColor, 'FontSize', fontSize);

% Display the image with text

imshow(imgWithText);

Answer 1

[Pratyush Swain]

Hi Merin,

I understand you utilized an image from the color thresholder app and performed ocr to extract text but did not get satisfactory results. Some steps that can be of help is further pre-processing of the input image before performing ocr to make sure the algorithm could read text properly.

You can refer to this page to know about image preprocessing techniques such as "morphological reconstruction","noise & background removal" , etc to improve ocr results: https://www.mathworks.com/help/vision/ug/recognize-text-using-optical-character-recognition-ocr.html.

Also another important point to note is OCR works best for an image containing dark text on a light background, so make sure to invert the binary image after exporting from color thresholding app & performing pre-processing steps:

% Assumin BW is binary image exported from app %
   % Perform Preprocessing steps ..... %
   
% Invert clean binarized image %
BW2 = imcomplement(BW);
figure
% Realize the difference between the two images %
imshowpair(BW,BW2,"montage")

% Finally perform ocr to extract text %
results = ocr(BW2);
results.Text

Also you can use the following text detection algorithms use local image features to locate or segment text within an image.

1- https://www.mathworks.com/help/vision/ug/automatically-detect-and-recognize-text-in-natural-images.html

2- https://www.mathworks.com/help/vision/ug/recognize-text-using-craft-model-and-ocr.html

For more information please refer to https://www.mathworks.com/help/vision/ug/getting-started-with-ocr.html.

Hope this helps.