forked from lightfield-analysis/evaluation-toolkit
-
Notifications
You must be signed in to change notification settings - Fork 0
/
SUBMISSION_INSTRUCTIONS
102 lines (80 loc) · 4.47 KB
/
SUBMISSION_INSTRUCTIONS
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
############################################################################
# This file is part of the 4D Light Field Benchmark. #
# #
# This work is licensed under the Creative Commons #
# Attribution-NonCommercial-ShareAlike 4.0 International License. #
# To view a copy of this license, #
# visit http://creativecommons.org/licenses/by-nc-sa/4.0/. #
# #
# Authors: Katrin Honauer & Ole Johannsen #
# Contact: [email protected] #
# Website: www.lightfield-analysis.net #
# #
# The 4D Light Field Benchmark was jointly created by the University of #
# Konstanz and the HCI at Heidelberg University. If you use any part of #
# the benchmark, please cite our paper "A dataset and evaluation #
# methodology for depth estimation on 4D light fields". Thanks! #
# #
# @inproceedings{honauer2016benchmark, #
# title={A dataset and evaluation methodology for depth estimation on #
# 4D light fields}, #
# author={Honauer, Katrin and Johannsen, Ole and Kondermann, Daniel #
# and Goldluecke, Bastian}, #
# booktitle={Asian Conference on Computer Vision}, #
# year={2016}, #
# organization={Springer} #
# } #
# #
# Version 1.0 #
############################################################################
INTRODUCTION
============
This file provides submission instructions for the 4D Light Field Benchmark.
The benchmark consists of 12 scenes:
- 4 stratified scenes (Backgammon, Pyramids, Dots, Stripes)
- 4 training scenes (Boxes, Cotton, Dino, Sideboard)
- 4 test scenes (Bedroom, Bicycle, Herbs, Origami)
You may use our evaluation toolkit to validate submissions and to evaluate
your method on the stratified and training scenes. Please check out the
README for further information on the software features.
SUBMISSION INSTRUCTIONS
=======================
1) Run your stereo algorithm.
INPUT:
Per scene, you may use the 81 light field views and the disparity
ranges as provided in the "parameters.cfg". You are not allowed to
use any of the provided evaluation masks for your depth estimation.
YOU ARE REQUESTED TO USE ONLY ONE SET OF PARAMETERS FOR ALL 12 SCENES.
OUTPUT:
Please report a 512x512 disparity map in PFM format and the runtime
in seconds for each of the 12 scenes. Have a look at the baseline
algorithms for examples of the expected file formats. We provide
Matlab and Python tools with PFM read/write utilities on our website.
Official PFM documentation: http://netpbm.sourceforge.net/doc/pfm.html
2) Finalize and validate your submission.
Create a zip file with the following file structure:
algo.zip
|-- disp_maps
|-- backgammon.pfm
|-- bedroom.pfm
|-- ...
|-- runtimes
|-- backgammon.txt
|-- bedroom.txt
|-- ...
You may use the "validate_submission.py" to check your submission.
3) Upload your results.
To submit your results and to manage your submissions, please
log in to our website: www.lightfield-analysis.net
Create a new submission, enter details about your method, and
upload your zip archive.
4) Check out your performance scores.
You will get notified via email as soon as your submission was
successfully evaluated and approved. Go to the benchmark
visualizations to check your scores and rankings.
CONTACT
========
In case you have any questions or discovered a bug, please contact us at:
For further information and recent updates, please visit:
www.lightfield-analysis.net