عرض عادي
عرض مارك
Face image analysis via Transformers: Algorithms and (رقم التسجيلة. 18346)
[ عرض عادي ]
| 100 1# - 100 | |
|---|---|
| a | Boukhari, Djamel Eddine |
| 245 00 - 245 | |
| a | Face image analysis via Transformers: Algorithms and |
| 260 ## - 260 | |
| b | Universite Chahid Hamma Lakhdar d'El-Oued |
| c | 2024 |
| 942 ## - 942 | |
| c | THESIS |
| 999 ## - 999 | |
| c | 18346 |
| d | 18346 |
| 952 ## - 952 | |
| 9 | 44867 |
| a | MAIN |
| b | MAIN |
| d | 2026-06-02 |
| o | TD621/041/01 |
| p | MAIN-1-16635 |
| y | THESIS |
| 500 ## - 500 | |
| -- | Face image analysis via Transformers: Algorithms and |
| 520 ## - 520 | |
| -- | The perception of beauty has long been a central topic in human society, shaped<br/>by socioeconomic, cultural, and historical influences. Despite the evolving opinions<br/>on facial beauty worldwide, understanding the factors behind facial attractiveness<br/>remains a key area of interest across disciplines such as psychology, computer<br/>science, and evolutionary biology. With advancements in computer vision and<br/>deep learning, facial beauty prediction (FBP) has emerged as a cutting-edge field,<br/>enabling objective quantification of facial beauty and its underlying factors.<br/>This thesis proposes four novel approaches to facial beauty prediction using<br/>deep learning. Two approaches leverage convolutional neural networks (CNNs)<br/>integrated with ensemble learning, combining predictions from multiple models<br/>to improve accuracy. The remaining two approaches harness the power of Vision<br/>Transformers, utilizing attention mechanisms to capture intricate relationships<br/>within facial features. Together, these methods enhance feature representation and<br/>analysis for robust and reliable facial beauty assessment.<br/>Experiments conducted on the SCUT-FBP5500 benchmark dataset demonstrate<br/>the effectiveness of our approaches. We achieve superior performance by comparing<br/>various deep learning models, including AlexNet, ResNet-18, and ResNeXt-50. The<br/>proposed models yield predictions that closely align with human evaluations, surpassing<br/>conventional methods in accuracy and consistency. This thesis underscores<br/>the transformative impact of deep learning in facial beauty prediction, offering<br/>precise, unbiased, and automated evaluations of facial attractiveness. |
| 650 #4 - 650 | |
| -- | /Face//image//analysis//via//Transformers//Algorithms//and/ |
| 700 1# - 700 | |
| -- | Chemsa, Ali |
لا توجد مواد متاحة.