Age and Gender Detection to Detect the Manipulated Images using CNN

There has been a proliferation of photographs during the last two decades. The number of photographs of human faces accessible has exploded in recent years, thanks in large part to the proliferation of smartphones and the rise in popularity of selfies. This has led to a surge in research into method...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:International Conference on Smart Systems and Inventive Technology (Online) s. 1187 - 1192
Hlavní autori: Chowdary, Boyilla Sushma, Subhadra, Valiveti Nagavalli, Kavitha, S.
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 23.01.2023
Predmet:
Benchmark testing
Convolution
Convolutional Neural Network
Convolutional neural networks
Deep Convolution Neural Network
Feature extraction
Java Database Connectivity
OpenCV
Principal Component Analysis
Social networking (online)
Supervised learning
Supervised Machine learning
Training data
ISSN:2832-3017
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:There has been a proliferation of photographs during the last two decades. The number of photographs of human faces accessible has exploded in recent years, thanks in large part to the proliferation of smartphones and the rise in popularity of selfies. This has led to a surge in research into methods for accurately determining a person age and gender solely from a photograph of the face. This research work intends to tackle this intricate challenge. Specifically, this study examines methods for determining a person's age, gender, and other characteristics based only on a static portrait of their face. Distinct models are trained to perform each task and compare the results of utilizing pre-trained CNN (Convolutional Neural Network) designs like VGG16 and ResNet50 and SE-ResNet50 on the VGGFace2 dataset to train the features collected by these networks. In addition to sharing the best practices to perform feature extraction using machine learning, this study also provides a benchmark performance analysis on several techniques. Even the most basic linear regression can be trained so that the extracted data performs better, however, CNNs were trained from scratch for age estimation.
ISSN:2832-3017
DOI:10.1109/ICSSIT55814.2023.10060905