-
Notifications
You must be signed in to change notification settings - Fork 45
/
demo_bt_ranking.m
executable file
·134 lines (103 loc) · 3.41 KB
/
demo_bt_ranking.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
% -------------------------------------------------------------------------
% Description:
% Demo script to calculate the BT scores
% This script reproduces the results of Figure 5 in our paper.
%
% Citation:
% A Comparative Study for Single Image Blind Deblurring
% Wei-Sheng Lai, Jia-Bin Huang, Zhe Hu, Narendra Ahuja, and Ming-Hsuan Yang
% IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016
%
% Contact:
% Wei-Sheng Lai
% University of California, Merced
% -------------------------------------------------------------------------
%% input dataset and attribute
dataset = 'real';
% dataset = 'uniform';
% dataset = 'nonuniform';
attribute = 'all';
% attribute = 'manmade';
% attribute = 'natural';
% attribute = 'people';
% attribute = 'saturated';
% attribute = 'text';
%% Load list
list_filename = fullfile('list', 'method.txt');
method = load_list(list_filename);
attr_filename = fullfile('attributes', sprintf('%s_%s.txt', dataset, attribute));
fprintf('Load %s\n', attr_filename);
attr_list = dlmread(attr_filename);
num_img = length(attr_list);
num_method = 14;
%% load votes
vote_filename = fullfile('votes', sprintf('votes_%s_balance_%s.csv', dataset, attribute));
fprintf('Load %s\n', vote_filename);
M = csvread(vote_filename, 1, 0); % offset the first row to skip header
%% build winning matrix
score_matrix = zeros(num_img, num_method);
for i = 1:num_img
row = find(M(:, 1) == attr_list(i));
% convert M to winning matrix
C = construct_winning_matrix(M(row, :), num_method);
% compute BT scores
score = BT_EM_exp(C);
% avoid Nan and Inf when compute mean
s = score;
s = s(s == s); % remove NaN
s = s(s ~= Inf); % remove Inf
% normalize to zero mean
score = score - mean(s);
score_matrix(i, :) = score';
end
%% compute mean scores for each method
method_score = zeros(num_method, 1);
for i = 1:num_method
s = score_matrix(:, i);
s = s(s == s); % remove NaN
s = s(s ~= Inf); % remove Inf
method_score(i) = mean(s);
end
%% sort methods by mean scores
[~, order] = sort(method_score, 'descend');
%% plot cumulative frequency
[color, marker, line_style] = color_spec;
figure; hold on;
select_method = order(1:14);
legend_cmd = 'legend(';
x = -5:0.5:10;
for i = 1:length(select_method)
m = select_method(i);
s = score_matrix(:, m);
s = s(s == s); % remove NaN
s = s(s ~= Inf); % remove Inf
y = histc(s, x);
z = cumsum(y)/sum(y);
h = plot(x, z, 'LineWidth', 2);
h.Color = color{m};
h.LineStyle = line_style{m};
legend_cmd = sprintf('%s method{%d}', legend_cmd, select_method(i));
if( i < length(select_method) )
legend_cmd = sprintf('%s, ', legend_cmd);
end
end
legend_cmd = sprintf('%s);', legend_cmd);
l = eval(legend_cmd);
l.FontSize = 16;
l.Location = 'southeast';
if( strcmp(attribute, 'manmade') )
title(sprintf('%s (man-made)', dataset));
else
title(sprintf('%s (%s)', dataset, attribute));
end
xlabel('B-T Scores');
ylabel('Cumulative Frequency');
h = gca;
h.FontName = 'Times New Roman';
h.FontSize = 24;
hold off;
%% save results
% filename = sprintf('bt_cdf_%s_%s.pdf', dataset, attribute);
% saveas(h, filename);
% fprintf('Save %s\n', filename);