bonus.m

%% Initialisation
%==========================================================================
% Add the path of used library.
% - The function of adding path of liblinear and vlfeat is included.
% - The use image directory is also included in this part.
% - image_dir{1} is the training positive face images(resize).
% - image_dir{2} is the training negative non-face images(resize).
% - val_dir is the validation set of real images
%==========================================================================
clear all
close all
clc
run ICV_setup
% The relevant data directory
images_dir{1} = './data/face_detection/cropped_faces/';     % positive samples directory
images_dir{2} = './data/face_detection/non_faces_images/';  % negative samples directory
face_images_dir = dir(images_dir{1});
face_images_dir(1:2)=[];

val_dir{1} = './data/face_detection/val_face_detection_images/'; % Validation data (For visualization purpose).
val_file = dir(val_dir{1});
val_file(1:2)=[];
val_dir{2} = './data/face_detection/te_raw_images/';            % Validation data (For performance evaluation)
val_file2 = dir(val_dir{2});
val_file2(1:2)=[];

% Hyperparameter of experiments
resize_size=[64 64];

% if ~exist('net')
% Setup MatConvNet.
addpath(genpath('./library/matconvnet/matlab'))
vl_setupnn();

% Load the VGG-Face model.
modelPath = fullfile(vl_rootnn,'data','models','vgg-face.mat') ;
if ~exist(modelPath)
  fprintf('Downloading the VGG-Face model ... this may take a while\n') ;
  mkdir(fileparts(modelPath)) ;
  urlwrite(...
    'http://www.vlfeat.org/matconvnet/models/vgg-face.mat', ...
    modelPath) ;
end

% Load the model and upgrade it to MatConvNet current version.
net = load(modelPath);
net = vl_simplenn_tidy(net);


%%

disp('Extracting features...')

hog  = true;
lbp  = true;

hog_cellSize   = 8;
lbp_cellSize   = 4;

max_resize = 1.0;
min_resize = 0.8;
threshold = 0.0;


%% Single/Multi-Scale Sliding Window
%==========================================================================
% Evaluating your detector and the sliding window.
% -It is okay to only use a single-scale sliding window for this assignment.
% However, a better performance would be required a multi-scale sliding
% window due to the different face size in real image.
%                                (You should finish this part by yourself)
%==========================================================================

load('face_detector.mat')

img = imread('test.png');
plt_img = img;
if size(img, 3)>1, img = rgb2gray(img); end
window_size = [64 64];

% Sliding window function
[patches, temp_bbox] = sw_detect_face(img, window_size, max_resize, 8);

for p = max_resize - 0.1:-0.1:min_resize
    [temp_patches, temp_bbox2] = sw_detect_face(img, window_size, p, 8);
    patches = cat(1, patches, temp_patches);
    temp_bbox = cat(1, temp_bbox, temp_bbox2);
end

% Extract the features for each patch
total = 0;

for p = 1:length(patches)
    for j = 1:size(patches{p}, 3)
        total = total + 1;
    end
end

v = resize_size(1) / hog_cellSize;
te_hog_vectors = zeros(total, v * v * 31);

hog_iter = 1;

for p = 1:length(patches)
    for j = 1:size(patches{p}, 3)
        temp = single(patches{p}(:, :, j))/255;
        temp = vl_hog(temp, hog_cellSize);
        te_hog_vectors(hog_iter, :) = temp(:)';
        hog_iter = hog_iter + 1;
    end
end

v = resize_size(1) / lbp_cellSize;
te_lbp_vectors = zeros(total, v * v * 58);

lbp_iter = 1;

for p = 1:length(patches)
    for j = 1:size(patches{p}, 3)
        temp = single(patches{p}(:, :, j))/255;
        temp = vl_lbp(temp, lbp_cellSize);
        te_lbp_vectors(lbp_iter, :) = temp(:)';
        lbp_iter = lbp_iter + 1;
    end
end

Xte = [];
Xte = cat(2, Xte, te_hog_vectors);
Xte = cat(2, Xte, te_lbp_vectors);

bbox_ms = [];

for p = 1:length(patches)
    for j = 1:size(patches{p}, 3)
        bbox_ms = [bbox_ms; temp_bbox{p}(j, :)];
    end
end

addpath('library/liblinear-2.1/windows/');

% Get the positive probability for proposed faces
[predicted_label, ~, prob_estimates] = predict(zeros(size(Xte, 1), 1), sparse(Xte), Mdl);
l = predicted_label;
score = prob_estimates;
prob2 = score(:, 1);

% Setting a threshold to pick the proposed face images
threshold_bbox = bbox_ms(prob2 > threshold, :);
prob3 = prob2(prob2 > threshold, :);

% Remove the redundant boxes via non-maximum supression.
% - The bbox is the top-left x, y, height, width of the patches.
% - prob2 is the confidence of the patches
[selectedBbox, selectedScore] = selectStrongestBbox(threshold_bbox, prob3, 'OverlapThreshold', 0.3, 'RatioType', 'Union');

% % Visualise the test images
% bbox_position = selectedBbox;
% figure
% imshow(plt_img)
% hold on
% for i = 1:size(bbox_position, 1)
% rectangle('Position', [bbox_position(i, 2),bbox_position(i, 1),bbox_position(i, 3:4)],...
%     'EdgeColor', 'b', 'LineWidth', 3)
% end

bbox_position = selectedBbox;
patches = zeros(size(selectedBbox, 1), resize_size(1), resize_size(2));

for i=1:size(selectedBbox, 1)
    x = bbox_position(i, 2);
    y = bbox_position(i, 1);
    window_x = bbox_position(i, 4);
    window_y = bbox_position(i, 3);
    
    patches(i, :, :) = imresize(plt_img(y:y+window_y-1, x:x+window_x-1), resize_size);
    
%     if size(plt_img(y:y+window_y-1, x:x+window_x-1), 1) ~= 64
%         patches(i, :, :) = padarray(plt_img(y:y+window_y-1, x:x+window_x-1), [0, 0]);
%     else
%         patches(i, :, :) = plt_img(y:y+window_y-1, x:x+window_x-1);
%     end
end

% x = bbox_position(2);
% y = bbox_position(1);
% window_x = bbox_position(4);
% window_y = bbox_position(3);
% 
% patches = plt_img(y:y+window_y-1, x:x+window_x-1);


%%

disp('Recognition: Extracting features...')

Xtr = []; Ytr = [];
Xva = []; Yva = [];

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Loading the training data
% -tr_img_sample/va_img_sample:
% The data is store in a N-by-3 cell array. The first dimension of the cell
% array is the cropped face images. The second dimension is the name of the
% image and the third dimension is the class label for each image.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

nn   = true;
pca_ = true;

pca_components = 250;


%% Part I: Face Recognition: Who is it?
%==========================================================================
% The aim of this task is to recognize the person in the image(who is he).
% We train a multiclass classifer to recognize who is the person in this
% image.
% - Propose the patches of the images
% - Recognize the person (multiclass)
%==========================================================================


disp('Recognition: Extracting features...')

Xva = []; 
Yva = [];

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Loading the training data
% -tr_img_sample/va_img_sample:
% The data is store in a N-by-3 cell array. The first dimension of the cell
% array is the cropped face images. The second dimension is the name of the
% image and the third dimension is the class label for each image.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

load('./models/fr_model.mat');

nn_vector_size = 2622;
va_nn_vectors = zeros(size(patches, 1), nn_vector_size);

for i = 1:size(patches, 1)
    temp = patches(i, :, :);
    temp = reshape(temp, [64,64]);
    temp = single(temp); % 255 range.
    temp = imresize(temp, net.meta.normalization.imageSize(1:2));
    temp = repmat(temp, [1, 1, 3]);
    temp = bsxfun(@minus, temp, net.meta.normalization.averageImage);
    temp = vl_simplenn(net, temp);
    temp = squeeze(temp(37).x);
    temp = temp./norm(temp,2);
    va_nn_vectors(i, :, :) = temp(:)';
end


%% Build data for training from extracted features
Xva = [Xva va_nn_vectors];

% PCA
Xva = bsxfun(@minus ,Xva, mean(Xva));
Xva = Xva * coeff;

Xva = double(Xva);

%% Train the recognizer and evaluate the performance

% model = train(double(Ytr), sparse(double(Xtr)));
[predicted_label, ~, prob_estimates] = predict(zeros(size(Xva, 1), 1), sparse(Xva), model);
l = predicted_label;
prob = prob_estimates;

database = ["Abdullah Gul", "Amelie Mauresmo", "Andy Roddick", "Angelina Jolie", "Atal Bihari", "Bill Gates", "Bill Simon", "Britney Spears", "Carlos Menem", "Carlos Moya", "David Nalbandian", "Dick Cheney", "Dominique de", "Eduardo Duhalde", "Fidel Castro", "George Robertson", "Halle Berry", "Hamid Karzai", "Hillary Clinton", "Hu Jintao", "Igor Ivanov", "James Blake", "Jean Charest", "Jennifer Aniston", "Jennifer Lopez", "Jeremy Greenstock", "Jiang Zemin", "John Bolton", "John Howard", "John Kerry", "John Snow", "Joschka Fischer", "Jose Maria", "Julianne Moore", "Julie Gerberding"];

% Visualise the test images
figure
imshow(plt_img)
hold on

for i = 1:size(bbox_position, 1)
    rectangle('Position', [bbox_position(i, 2),bbox_position(i, 1),bbox_position(i, 3:4)],...
        'EdgeColor', 'b', 'LineWidth', 3)
    text(double(bbox_position(i, 2))+5, double(bbox_position(i, 1))-15, database(l(i)), 'FontSize', 7, 'Color', 'y')
end