-
Notifications
You must be signed in to change notification settings - Fork 4
Expand file tree
/
Copy pathspecification.html
More file actions
95 lines (74 loc) · 5.77 KB
/
Copy pathspecification.html
File metadata and controls
95 lines (74 loc) · 5.77 KB
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
<!DOCTYPE html>
<html lang="en"><head>
</head>
<body>
<ul id="content_listContainer" class="contentList">
<li id="contentListItem:_942680_1" class="clearfix read">
<img alt="Assignment" src="/images/ci/sets/set01/assignment_on.gif" class="item_icon">
<div class="item clearfix" id="_942680_1">
<h3 id="anonymous_element_8">
<span class="reorder editmode hideme"><span><img src="/images/ci/icons/generic_updown.gif" alt=""></span></span>
<a href="/webapps/assignment/uploadAssignment?content_id=_942680_1&course_id=_38755_1&group_id=&mode=view"><span style="color:#000000;">Locate and Recognise Paintings</span></a>
</h3>
</div>
<div class="details">
<div class="vtbegenerated">
<p>Develop a program (in C++ using OpenCV) to locate and recognise painting in a gallery (e.g. the image below shows a number of painting in the gallery with the ideal answers shown in green). </p>
<p> <img width="1296" height="968" alt="" src="https://tcd.blackboard.com/courses/1/CS4053-A-Y-201718/content/_942680_1/embedded/G1GT2.jpg"></p>
<p>Your code must also automatically evaluate the performance of your system in terms of the Accuracy, Precision and Recall and any other metrics you feel are appropriate.</p>
<p><strong><span style="text-decoration: underline;">Project files:</span></strong></p>
<p>There are six known paintings and four gallery images in which these painting must be located. The images are available at <a onclick="this.href='/webapps/blackboard/content/contentWrapper.jsp?content_id=_942680_1&displayName=Linked+File&navItem=content&attachment=true&course_id=_38755_1&tab_group=courses&href=https%3A%2F%2Fwww.scss.tcd.ie%2Fpublications%2Fbook-supplements%2FA-Practical-Introduction-to-Computer-Vision-with-OpenCV%2FProblems%2FPaintings.zip';" href="https://www.scss.tcd.ie/publications/book-supplements/A-Practical-Introduction-to-Computer-Vision-with-OpenCV/Problems/Paintings.zip">https://www.scss.tcd.ie/publications/book-supplements/A-Practical-Introduction-to-Computer-Vision-with-OpenCV/Problems/Paintings.zip</a></p>
<p><strong><span style="text-decoration: underline;">Ground Truth:</span></strong></p>
<p>The ideal answers the identities and locations of the paintings as follows:</p>
<ul>
<li>Gallery 1:
<ul>
<li>Painting2 at (212, 261), (445, 225), (428, 725), (198, 673)</li>
<li>Painting1 at (686, 377), (1050, 361), (1048, 705), (686, 652)</li>
</ul> </li>
<li>Gallery 2:
<ul>
<li>Painting3 at (252, 279), (691, 336), (695, 662), (258, 758)</li>
<li>Painting2 at (897, 173), (1063, 234), (1079, 672), (917, 739)</li>
<li>Painting1 at (1174, 388), (1221, 395), (1216, 544), (1168, 555)</li>
</ul> </li>
<li>Gallery 3:
<ul>
<li>Painting4 at (68, 329), (350, 337), (351, 545), (75, 558)</li>
<li>Painting5 at (629, 346), (877, 350), (873, 517), (627, 530)</li>
<li>Painting6 at (1057, 370), (1187, 374), (1182, 487), (1053,493)</li>
</ul> </li>
<li>Gallery 4:
<ul>
<li>Painting4 at (176, 348), (298, 347), (307, 481), (184,475)</li>
<li>Painting5 at (469, 343), (690, 338), (692, 495), (472, 487)</li>
<li>Painting6 at (924, 349), (1161, 344), (1156, 495), (924, 488)</li>
</ul> </li>
</ul>
<p><strong><span style="text-decoration: underline;">Warning:</span></strong></p>
<p>This exercise is to be done on your own (no collaboration whatsoever is permitted).</p>
<p><strong><span style="text-decoration: underline;">Submission:</span></strong></p>
<p>You should submit this assignment by the due date. All submissions MUST be made through Blackboard AND MUST NOT BE IN ARCHIVE FILES (e.g. ZIP files)). You should submit</p>
<ul>
<li>a report in PDF format containing</li>
<li>a detailed description of your solution (i.e. how it works) including images to illustrate the processing where appropriate.</li>
<li>discussion of your solutions, any issues/problems (and suggestions for improvements if possible).</li>
<li>result images for all four test images showing the ground truth and your detected text regions.</li>
<li>the performance metrics (Precision, Recall, Accuracy and anything else deemed appropriate).</li>
<li>discssion of the results including any issues/problems with the metrics and the ground truth (and suggestions for improvements if possible)</li>
<li>the full source code as a separate attachment. (The code must be well written (easy to read and follow)).</li>
</ul>
<p><strong><span style="text-decoration: underline;">Marking Scheme:</span></strong></p>
<p>This lab will be graded based primarily on the submitted report. The code will be checked if there is any query over your implementation. The marks for this assignment will be awarded as follows:</p>
<p>40% Technique. Your solution to the problem must use well defined Computer Vision operations and be clearly explained.</p>
<p>10% Understanding of your solution. This is based on your discussion of the appropriateness of your solution and any issues relating to it's reliability on the test images or any other images (e.g. do you have any idea WHEN your solution will fail?).</p>
<p>40% Performance (i.e. how successful was your solution? All four test images with ground truth and your results must be included, as must the various required performance metrics. If these are missing performance marks will be reduced proportionately.) </p>
<p>10% Understanding of the metrics and the ground truth. How well does your discussion of the metrics and the ground truth show your understanding of the meaning of the results and their limitations...</p>
<br>
</div>
</div>
<div id="_942680_1previewModule" class="moduleSample">
</div>
</li>
</ul>
</body></html>