codeold.m

hopfield neural network for binary image recognition

function code

    % Read the sample image in
    im = imread('shapessm.jpg');
    
    % Find edges using the Canny operator with hysteresis thresholds of 0.1
    % and 0.2 with smoothing parameter sigma set to 1.
    edgeim = edge(im,'canny', [0.1 0.2], 1);

    figure(1), imshow(edgeim); truesize(1)
    
    % Link edge pixels together into lists of sequential edge points, one
    % list for each edge contour.  Discard contours less than 10 pixels long.
    [edgelist, labelededgeim] = edgelink(edgeim, 10);
    
    % Display the labeled edge image with separate colours for each
    % distinct edge (choose your favorite colourmap!)
    figure(2), imagesc(labelededgeim); colormap(vga), 
    axis image, axis off, truesize(2)
    
    % Fit line segments to the edgelists with the following parameters:
    tol = 2;         % Line segments are fitted with maximum deviation from
		     % original edge of 2 pixels.
    angtol = 0.05;   % Segments differing in angle by less than 0.05 radians
    linkrad = 2;     % and end points within 2 pixels will be merged.
    [seglist, nedgelist] = lineseg(edgelist, tol, angtol, linkrad);

    % Draw the fitted line segments stored in seglist in figure window 3 with
    % a linewidth of 2
    drawseg(seglist, 3, 2); axis off

%    figure(1), print -Pjpg -r0 edgeim.jpg
%    figure(2), print -Pjpg -r0 labeledgeim.jpg
%    figure(3), print -Pjpg -r0 segmentim.jpg