Exploring Deep Learning Approaches for Effective Gender Identification in Face Images

Authors

Abhishek Gupta, Prof. Vikas Kamle

Abstract

Computer vision extensively investigates the field of automatic gender determination from face photographs. Although humans find this activity rather straightforward, it presents a significant obstacle for robots. This article is a study that proposes a strategy for classifying gender based on feature vectors obtained from facial recognition. The procedure starts by using face recognition and preprocessing techniques on incoming photos, transforming them into a standardized format. Subsequently, a face recognition model is used to extract feature vectors that accurately describe the facial traits inside the designated feature space. Ultimately, the feature vectors are categorized using machine learning methodologies.
This paper presents a sophisticated approach for classifying gender using VGG Face and Deep Belief Networks, which includes adding shifted filter responses. The study examines many models, such as CNN, VGG 16, ResNet50, Inception v3, and EfficientNet, and finds that ResNet152 has greater performance. The ResNet152 model has superior performance compared to other models, with an estimated 9% improvement. This is attributed to its enhanced ability to handle outliers, surpassing the capabilities of earlier models.

Keywords

Gender classification, convolutional neural systems , SVM , LFW, FERET, CNN, VGG 16, Resnet50, inception v3, and EfficientNet

Exploring Deep Learning Approaches for Effective Gender Identification in Face Images. Abhishek Gupta, Prof. Vikas Kamle. 2024. IJIRCT, Volume 10, Issue 4. Pages 1-9. https://www.ijirct.org/viewPaper.php?paperId=2408067