-
Notifications
You must be signed in to change notification settings - Fork 119
/
evaluate_MSLapSRN_dataset.m
123 lines (93 loc) · 3.43 KB
/
evaluate_MSLapSRN_dataset.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
% -------------------------------------------------------------------------
% Description:
% Script to evaluate pretrained MS-LapSRN on benchmark datasets
%
% Citation:
% Fast and Accurate Image Super-Resolution with Deep Laplacian Pyramid Networks
% Wei-Sheng Lai, Jia-Bin Huang, Narendra Ahuja, and Ming-Hsuan Yang
% arXiv, 2017
%
% Contact:
% Wei-Sheng Lai
% University of California, Merced
% -------------------------------------------------------------------------
%% testing options
% dataset = 'Set5';
% dataset = 'Set14';
dataset = 'BSDS100';
% dataset = 'Urban100';
% dataset = 'Manga109';
% model_name = 'MSLapSRN_D5R2';
% model_name = 'MSLapSRN_D5R5';
model_name = 'MSLapSRN_D5R8';
model_scale = 2; % pretrained model upsampling scale
test_scale = model_scale; % testing scale can be different from model scale
gpu = 1; % GPU ID, gpu = 0 for CPU mode
compute_ifc = 0; % IFC calculation is slow, enable when needed
%% setup paths
input_dir = fullfile('datasets', dataset);
output_dir = fullfile('results', dataset, sprintf('x%d', test_scale), model_name);
if( ~exist(output_dir, 'dir') )
mkdir(output_dir);
end
addpath(genpath('utils'));
addpath(fullfile(pwd, 'matconvnet/matlab'));
vl_setupnn;
%% Load pretrained multi-scale model
model_filename = fullfile('pretrained_models', sprintf('%s.mat', model_name));
fprintf('Load %s\n', model_filename);
net = load(model_filename);
net_trained = dagnn.DagNN.loadobj(net.net);
opts_filename = fullfile('pretrained_models', sprintf('%s_opts.mat', model_name));
fprintf('Load %s\n', opts_filename);
opts = load(opts_filename);
opts = opts.opts;
opts.scales = [model_scale];
%% create single-scale model
net = init_MSLapSRN_model(opts, 'test');
%% copy pretrained weights
fprintf('Copy weights to single scale model...\n');
net = copy_model_weights(net, net_trained);
if( gpu ~= 0 )
gpuDevice(gpu)
net.move('gpu');
end
%% load image list
list_filename = fullfile('lists', sprintf('%s.txt', dataset));
img_list = load_list(list_filename);
num_img = length(img_list);
%% testing
PSNR = zeros(num_img, 1);
SSIM = zeros(num_img, 1);
IFC = zeros(num_img, 1);
for i = 1:num_img
img_name = img_list{i};
fprintf('Testing %s on %s %dx: %d/%d: %s\n', model_name, dataset, test_scale, i, num_img, img_name);
%% Load GT image
input_filename = fullfile(input_dir, sprintf('%s.png', img_name));
img_GT = im2double(imread(input_filename));
img_GT = mod_crop(img_GT, test_scale);
%% generate LR image
img_LR = imresize(img_GT, 1/test_scale, 'bicubic');
%% apply SR
img_HR = SR_MSLapSRN(img_LR, net, model_scale, test_scale, gpu);
%% save result
output_filename = fullfile(output_dir, sprintf('%s.png', img_name));
fprintf('Save %s\n', output_filename);
imwrite(img_HR, output_filename);
%% evaluate
[PSNR(i), SSIM(i), IFC(i)] = evaluate_SR(img_GT, img_HR, test_scale, compute_ifc);
end
PSNR(end+1) = mean(PSNR);
SSIM(end+1) = mean(SSIM);
IFC(end+1) = mean(IFC);
fprintf('Average PSNR = %f\n', PSNR(end));
fprintf('Average SSIM = %f\n', SSIM(end));
fprintf('Average IFC = %f\n', IFC(end));
filename = fullfile(output_dir, 'PSNR.txt');
save_matrix(PSNR, filename);
filename = fullfile(output_dir, 'SSIM.txt');
save_matrix(SSIM, filename);
filename = fullfile(output_dir, 'IFC.txt');
save_matrix(IFC, filename);