In this paper, we propose a novel facial wrinkle segmentation method based on the Unet++ model, enhanced with dice and focal loss functions. Our approach begins with the construction of an enriched wrinkle dataset sourced from the Flickr-Faces-HQ dataset, ensuring diversity in wrinkle types and complexities.
Nội dung trích xuất từ tài liệu:
A deep learning approach for accurate facial wrinkle segmentation using unet++ model with dice and focal loss functions
A deep learning approach for accurate facial wrinkle segmentation using unet++ model with dice and focal loss functions
Số trang: 7
Loại file: pdf
Dung lượng: 675.27 KB
Lượt xem: 1
Lượt tải: 0
Thông tin tài liệu:
Tìm kiếm theo từ khóa liên quan:
Deep learning Dice loss Facial wrinkle segmentation Focal loss Unet++ modelTài liệu có liên quan:
-
8 trang 235 0 0
-
Application of convolutional neural network for detecting concrete cracks
4 trang 44 0 0 -
Improving hand posture recognition performance using multi-modalities
10 trang 40 0 0 -
11 trang 38 0 0
-
Modern approaches in natural language processing
25 trang 38 0 0 -
8 trang 37 0 0
-
91 trang 35 0 0
-
8 trang 35 0 0
-
Research on traffic congestion detection from camera images in a location of Da Lat
13 trang 34 0 0 -
7 trang 33 0 0
