Separate strokes from a gesture

I have to separate individual strokes from a gesture like curved line and straight line from numeral "2". How to do it?

Answer 1

[Pratyush]

Hi Shweta,

To separate individual strokes from a gesture like the numeral "2" in MATLAB, follow these steps:

• Pre-process the image by converting it to grayscale, binarizing it, and removing noise.
• Use edge detection and skeletonization to find and thin the outlines of the strokes.
• Trace the contours of the strokes and analyze their curvature to distinguish between straight and curved segments.
• Label and classify the different strokes based on features such as length and curvature.
• Visualize the separated strokes for verification.

Here's an example MATLAB code for these steps:

% Load and preprocess the image
img = imread('numeral2.png');
grayImg = rgb2gray(img);
binaryImg = imbinarize(grayImg);
cleanImg = bwareaopen(binaryImg, 30);

% Edge detection and skeletonization
edges = edge(cleanImg, 'canny');
skeleton = bwmorph(edges, 'thin', Inf);

% Find contours and separate strokes
[contours, ~] = bwboundaries(skeleton, 'noholes');
for i = 1:length(contours)
    contour = contours{i};
    % Analyze contour for curvature and classify (code for analysis not included)
    % Visualize the contour
    figure; imshow(skeleton); hold on;
    plot(contour(:,2), contour(:,1), 'g', 'LineWidth', 2);
end

The actual curvature analysis and classification logic are not included and would need to be implemented based on your specific requirements.