-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
128 lines (103 loc) · 3.9 KB
/
index.html
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
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<meta name="description" content="Dynamic Head Reconstruction Project">
<meta name="keywords" content="3D Head Reconstruction, Identity, Expression, Head Pose, Deep Learning">
<meta name="author" content="Levon Avetisyan">
<title>Dynamic Head Reconstruction Project</title>
<style>
body {
font-family: 'Arial', sans-serif;
background-color: #f4f4f4;
margin: 0;
padding: 0;
}
header {
background-color: #343a40;
color: white;
padding: 1em;
text-align: center;
}
h1 {
font-size: 2.5em;
margin: 0;
}
main {
max-width: 1000px;
margin: 2em auto;
padding: 0 2em;
}
section {
background-color: white;
border-radius: 8px;
padding: 20px;
margin-bottom: 1em;
box-shadow: 0px 2px 4px rgba(0, 0, 0, 0.1);
}
h2 {
margin-top: 0;
}
.center {
text-align: center;
}
.btn-container {
margin-top: 1.5em;
}
.btn {
display: inline-block;
background-color: #343a40;
color: white;
padding: 0.5em 1.5em;
border-radius: 8px;
text-decoration: none;
margin: 0.5em;
font-weight: bold;
}
.btn:hover {
background-color: #55595c;
}
.abstract {
font-size: 1.1em;
line-height: 1.6;
}
.video-embed {
max-width: 100%;
height: auto;
margin: 1.5em 0;
}
.footer {
text-align: center;
margin-top: 2em;
padding: 1em;
background-color: #343a40;
color: white;
}
</style>
</head>
<body>
<header>
<h1>Reconstructing 3D Faces From 2D Inputs</h1>
</header>
<main>
<section class="center">
<h2>Abstract</h2>
<p class="abstract">
This thesis explores the reconstruction of identity, expression, and head pose from 2D images, utilizing advanced deep learning models. The primary objective is to disentangle facial identity and expression while estimating accurate head pose transformations. The research focuses on the Famudy dataset, which provides detailed annotations for facial identities, expressions based on the NPHM model, and head pose data.
A range of pretrained convolutional neural networks were adapted for the expression and identity reconstruction tasks. Custom loss functions and data augmentation strategies were employed to enhance model performance and generalization, particularly considering dataset’s size limitations.
The experiments demonstrate strong results for expression code reconstruction, though identity and head pose reconstruction remained challenging due to the limited number of unique identities in the dataset.
This thesis provides a detailed evaluation of the models used, outlining the strengths and limitations of each approach. The findings of this thesis, offer valuable insights into facial reconstruction and head pose estimation, which can become a base for future explorations and be applied in various fields.
</p>
</section>
<section class="center">
<h2>Project Links</h2>
<div class="btn-container">
<a href="https://arxiv.org/abs/1234567" class="btn">Paper</a>
<a href="https://github.com/YourGitHubRepo" class="btn">Code</a>
<a href="https://your-project-video-link" class="btn">Video</a>
</div>
</section>
</main>
</body>
</html>