I want to give a piece of code an object, and have it find what the closest thing to that object is

So the current problem that I'm dealing with is that I have images where cells are compartmentalized in a channel, and I need to identify where this channel actually begins. This is not a machine learning image recognition problem, there is a channel in every photo -- I need the code to essentially, given what the channel looks like, try to find which part of the photo is most likely the channel, and return the position. Anyone got any clues on how I could potentially solve this? I realize this is a regression problem more or less, just wondering if there is a more elegant solution.

[Walter Roberson]

What characterizes a "channel" for this purpose?

[andi shahaj]

Geometric shape really, looks like a pill, basically a long rectangle capped by semi circles on either side. I should mention that usually only half of this channel is visible, so the defining feature is that semicircle

[Walter Roberson]

Does it have a particular orientation? Does it have a particular length? Does it have a particular aspect ratio? Are there other similar objects (such as semi-circles) in the image that are not part of a channel? Is there a color difference? Is there a texture difference? Is there a sample image we could look at?

[andi shahaj]

It has the same orientation every time (or almost the same, change of less than one degree), the aspect ration is a little variable but should not change too drastically, there is a colour difference the channel is much brighter, the texture is different as well. I'll attach a picture here for you

[andi shahaj]

See that rounded bit on the right? That's the part I want to identify.

Answer 1

[Image Analyst]

Try this. Adjust the threshold to widen or narrow the detected region.

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 = 25;

%===============================================================================
% Get the name of the image the user wants to use.
baseFileName = 'Screen Shot 2017-08-08 at 5.24.20 PM.png';
% Get the full filename, with path prepended.
folder = pwd
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

%===============================================================================
% Read in demo image.
rgbImage = imread(fullFileName);
% Get the dimensions of the image.
[rows, columns, numberOfColorChannels] = size(rgbImage);
% Display the original image.
subplot(2, 2, 1);
imshow(rgbImage, []);
axis on;
caption = sprintf('Original Color Image, %s', baseFileName);
title(caption, 'FontSize', fontSize, 'Interpreter', 'None');
hp = impixelinfo();

% Set up figure properties:
% Enlarge figure to full screen.
set(gcf, 'Units', 'Normalized', 'OuterPosition', [0 0.05 1 0.95]);
% Get rid of tool bar and pulldown menus that are along top of figure.
% set(gcf, 'Toolbar', 'none', 'Menu', 'none');
% Give a name to the title bar.
set(gcf, 'Name', 'Demo by ImageAnalyst', 'NumberTitle', 'Off')

drawnow;
hp = impixelinfo(); % Set up status line to see values when you mouse over the image.

% Extract the individual red, green, and blue color channels.
redChannel = rgbImage(:, :, 1);
% greenChannel = rgbImage(:, :, 2);
% blueChannel = rgbImage(:, :, 3);

% Sum the red channel vertically to get the horizontal profile.
horizontalProfile = mean(redChannel, 1);

% Plot the profile.
subplot(2, 2, 2);
plot(horizontalProfile, 'r-', 'LineWidth', 2);
grid on;
title('Horizontal Profile', 'FontSize', fontSize, 'Interpreter', 'None');

% Find out where the profile exceeds some threshold.
threshold = 95;
aboveThreshold = horizontalProfile > threshold;
% Extract just the longest region.
aboveThreshold = bwareafilt(aboveThreshold, 1);
% Find the left-most column
startingColumn = find(aboveThreshold, 1, 'first')
% Find the right-most column
endingColumn = find(aboveThreshold, 1, 'last')

% Display the image again.
% Display the original image.
subplot(2, 2, 3:4);
imshow(rgbImage);
axis on;
caption = sprintf('Original Color Image, with starting and ending lines');
title(caption, 'FontSize', fontSize, 'Interpreter', 'None');
hp = impixelinfo();
% Draw line over it at the starting column
hold on;
line([startingColumn, startingColumn], [1, rows], 'Color', 'b', 'LineWidth', 3);
line([endingColumn, endingColumn], [1, rows], 'Color', 'b', 'LineWidth', 3);

% Draw lines over profile plot.
subplot(2, 2, 2);
hold on;
line([startingColumn, startingColumn], ylim, 'Color', 'b', 'LineWidth', 3);
line([endingColumn, endingColumn], ylim, 'Color', 'b', 'LineWidth', 3);

[andi shahaj]

Bless your kind heart, once again you have saved me :)