VGU RESEARCH REPOSITORY
Please use this identifier to cite or link to this item:
https://epub.vgu.edu.vn/handle/dlibvgu/1284
Title: | BMI prediction from facial images using deep neural network | Authors: | Pham Quoc Trung | Keywords: | BMI prediction;Facial images;Deep neural network | Issue Date: | 2021 | Abstract: | According to many studies, face images can provide quite a lot of information about health status. The outburst of social networks and mobile devices allows researchers to access this information more accessible for many health care purposes. This project focuses on estimating Body mass index (BMI) based on facial images using the deep neural network because of its outstanding performance in many computer vision tasks. Specifically, we adopt the concept proposed by Dantcheva, Bremond and Bilinski, 2018 and deploy an efficient neural network, namely ResNet, as features extractor, followed by fully connected layers in charge of regression model for BMI estimation. After being trained on the VIP Attribute dataset, we evaluated this model for accuracy on two distinct datasets. The initial dataset randomly is chosen 20% from the VIP Attribute set, and the model’s mean absolute error (MAE) reached 2.8. For the dataset collected from Vietnamese participants, the MAE is 2.57. The positive result demonstrates the potential of creating a large-scale computer system for BMI prediction using facial pictures |
URI(1): | https://epub.vgu.edu.vn/handle/dlibvgu/1284 |
Appears in Collections: | Computer Science (CS) |
Files in This Item:
File | Description | Size | Format | Existing users please Login |
---|---|---|---|---|
BMI prediction from facial images using deep neural network.pdf | 3.15 MB | Adobe PDF |
Page view(s)
49
checked on Apr 17, 2024
Download(s)
21
checked on Apr 17, 2024
Google ScholarTM
Check
Items are protected by © Copyright of Vietnamese - German University Library