GENDER & AGE DETECTION USING GEOMETRIC-BASED & APPEARANCE-BASED DEEP LEARNING APPROACH ON REAL-TIME VIDEO CROWD

Main Article Content

Dimpal A. Kapgate, Prof. Abhimanyu Dutonde, Prof. Pragati Patil

Abstract

Abstract


Gender Classification method using biometric features, we access from the face of a human. That means the face of the human being is an important biometric feature. As well as Age Estimation is also a challenging task in object recognition.  Due to differences in head attitude, scale fluctuation in face photos, varying lighting conditions, occluded faces, and noisy face images, identification of faces and classification according to gender have become extremely challenging tasks. Two feature extraction techniques have been used: appearance-based and geometric-based techniques. Variations in lighting circumstances, head posture, facial expressions, partial occlusion by hats or spectacles, and camera quality can all affect how well the gender classification algorithm performs in terms of categorization rate. Therefore, it is ideal to have an algorithm that can withstand changes in illumination, position, occlusion, and emotion. The study paper's primary goal is to examine an automatic gender detection technique employing Deep Convolution Neural Networks (D-CNNS) on facial photos. The study paper's second goal is to evaluate an algorithm for age detection utilizing a novel geometric-based approach and appearance-based method based on the image, taking into account factors such as significant illumination variance. This research paper analyzes gender categorization for computer vision applications and presents improvements in gender classification & age detection accuracy. It has been discovered that texture descriptors perform a better job of classifying gender & Age detection than edge descriptors. The classification of gender that combines more characteristics provides greater accuracy than the other methods covered in this study. The intensity, shape, and texture of the face image are the features that are employed to get the best accuracy. As a result, the research project offers a thorough analysis of gender classification & Age Detection with improved accuracy when compared to the body of existing literature. Future studies in this area may involve developing a novel method for classifying gender in addition to identifying the Transgender ethnicity of voice captured in unrestricted environments.


 

Article Details

How to Cite
Prof. Pragati Patil, D. A. K. P. A. D. (2024). GENDER & AGE DETECTION USING GEOMETRIC-BASED & APPEARANCE-BASED DEEP LEARNING APPROACH ON REAL-TIME VIDEO CROWD. Obstetrics and Gynaecology Forum, 34(3s), 2633–2644. Retrieved from https://obstetricsandgynaecologyforum.com/index.php/ogf/article/view/794
Section
Articles

Similar Articles

1 2 3 4 5 6 7 8 > >> 

You may also start an advanced similarity search for this article.