-
Notifications
You must be signed in to change notification settings - Fork 2
/
run_example.m
39 lines (31 loc) · 1.14 KB
/
run_example.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
clear all; close all; clc;
%%
% "Seeing Around Corners with Edge-Resolved Transient Imaging"
%
% Joshua Rapp, Charles Saunders, Julian Tachella, John Murray-Bruce,
% Yoann Altmann, Jean-Yves Tourneret, Stephen McLaughlin,
% Robin M. A. Dawson, Franco N. C. Wong & Vivek K Goyal
%
% Nature Communications
% Contact: [email protected]
%% Choose a dataset to process
% Hidden scene name. Options below:
% Real data: 'mannequin', 'mannequins', 'empty_room',
% 'big_empty_room', 'board', 'staircase'
% Synthetic data: 'synthetic_verticalT', 'synthetic_farT_cylinder',
% 'synthetic_farT', 'synthetic_empty_room', 'synthetic_cylinder'
scene = 'staircase';
% Acquisition time in seconds. Only for real data.
% Options: 10,20,30 and 60.
acq_time = 30;
%% Run Skellapop algorithm
plot_debug = false; % boolean enabling intermediate plotting of the MCMC estimation
MCMC_iterations = 3e3; % number of MCMC iterations
% other hyperparameters can be tuned in get_hyperparam.m
addpath('skellapop_fcns');
if contains(scene, 'synthetic')
filename = scene;
else
filename = [scene '_' num2str(acq_time) 's'];
end
skellapop(filename, MCMC_iterations, plot_debug);