main.m

% shady nikooei

function main
    clc;
    clear;
    close all;

    reference_folder = 'E:\shady\term6\Image Process\06_Nikooei_WB_2_403\ex22\source_image';      
    query_image_path = '6.jpg';
    
    % feature weights for the 8-element feature vector
    % [Area, Eccentricity, AspectRatio, Solidity, Perimeter, Circularity, Extent, EulerNumber]
    weights = [1, 2, 2, 1, 1, 3, 2, 1]; 

    % build the reference feature library
    image_files = dir(fullfile(reference_folder, '*.jpg'));
    if isempty(image_files)
        error('No reference images found in the specified folder: %s', reference_folder);
    end
    
    % preallocation 
    num_references = length(image_files);
    ref_features = zeros(num_references, 8); % Updated to hold 8 features
    labels = cell(1, num_references);
    
    % save name of source object
    for i = 1:num_references
        current_image_path = fullfile(reference_folder, image_files(i).name);
        [~, filename, ~] = fileparts(image_files(i).name);
        labels{i} = filename;
                
        % save extracted features
        img = imread(current_image_path);
        ref_features(i, :) = extract_shape_features(img);
    end

    q_img = imread(query_image_path);
    b_qimg = preprocess_image(q_img);
    
    % extract all properties of query objects
    properties_to_extract = {'Area', 'BoundingBox', 'Eccentricity', 'Solidity', 'Centroid', 'Perimeter', 'Circularity', 'Extent', 'EulerNumber'};
    q_info = regionprops(b_qimg, properties_to_extract);

    % classify and display results
    figure;
    imshow(q_img);
    title('Object Detection Results (Enhanced Features)');
    hold on;
    for i = 1:length(q_info)

        % extract all 8 features for the current object
        area_i = q_info(i).Area;
        ecc_i = q_info(i).Eccentricity;
        asp_i = q_info(i).BoundingBox(3) / q_info(i).BoundingBox(4);
        sol_i = q_info(i).Solidity;
        perim_i = q_info(i).Perimeter;
        circ_i = q_info(i).Circularity;
        ext_i = q_info(i).Extent;
        eul_i = q_info(i).EulerNumber;
        
        obj_feat = [area_i, ecc_i, asp_i, sol_i, perim_i, circ_i, ext_i, eul_i];
        
        % compare with reference library
        errors = abs(ref_features - obj_feat);
        
        % normalize size-dependent features (Area and Perimeter)
        errors(:, 1) = errors(:, 1) / max(ref_features(:, 1));
        errors(:, 5) = errors(:, 5) / max(ref_features(:, 5));
        
        score = sum(errors .* weights, 2);
        
        [~, best_match_idx] = min(score);
        detected_label = labels{best_match_idx};
        
        % display the label and bounding box
        rectangle('Position', q_info(i).BoundingBox, 'EdgeColor', 'r', 'LineWidth', 2);
        centroid = q_info(i).Centroid;
        text(centroid(1), centroid(2), detected_label, ...
             'Color', 'cyan', 'FontSize', 14, 'FontWeight', 'bold', ...
             'HorizontalAlignment', 'center', 'BackgroundColor', 'black');
    end
    hold off;
end