NORMA eResearch @NCI Library

Human Face Analysis using Transfer Learning Approach

Singh, Gaurav (2023) Human Face Analysis using Transfer Learning Approach. Masters thesis, Dublin, National College of Ireland.

[thumbnail of Master of Science]
Preview
PDF (Master of Science)
Download (6MB) | Preview
[thumbnail of Configuration manual]
Preview
PDF (Configuration manual)
Download (10MB) | Preview

Abstract

The person’s face is among the most significant body parts for transmitting key features since it is so prominently shown. People of varying ages, sexes, and races each have a unique set of facial characteristics. Numerous studies have indicated that smokers have a significantly increased risk of experiencing substantial alterations in the middle and lower thirds of their faces. This research investigates the use of five distinct pre-trained - DCNN(deep convolutional neural network) learning models for identifying human face traits by making use of three different datasets and doing a comparative analysis. It is possible for governments, schools, insurance companies, and universities to leverage automation of large-scale BMI imputations, as well as age, gender, ethnicity, and smoking status, in order to collect data on a range of social concerns and make more informed decisions. Multiple studies have been done, with each one concentrating on a limited number of attributes and making use of a more condensed data set. In this particular investigation, pre-defined transfer learning networks such as EfficientNetB7, DenseNet201, InceptionResNetV2, VGG16, and ResNet50V2 were chosen over conventional-CNN (convolutional neural networks) in order to compare the results of the pre-trained models with one another. This comparison was carried out in order to determine which of the models performed the best. This was accomplished by employing a two-way comparison strategy. This study aims to quantify essential human face attributes and evaluate the findings based on parameters such as precision, accuracy, recall, RMSE, and MAE using a variety of methods that are considered to be state-of-the-art in the field.

Item Type: Thesis (Masters)
Supervisors:
Name
Email
Basilio, Jorge
UNSPECIFIED
Subjects: Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science
R Medicine > RA Public aspects of medicine
T Technology > TK Electrical engineering. Electronics. Nuclear engineering > Biometric Identification
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: Tamara Malone
Date Deposited: 26 May 2023 13:29
Last Modified: 26 May 2023 13:29
URI: https://norma.ncirl.ie/id/eprint/6665

Actions (login required)

View Item View Item