Automated Deep Learning based Age and Gender Classification Model using Facial Features for Video Surveillance

Main Article Content

S. Sivachandiran, Dr. K. Jagan Mohan, Dr. G. Mohammed Nazer


Video surveillance plays a vital role in ensuring security in public and private places over the globe. Facial image analysis using computer vision (CV) and artificial intelligence (AI) based tools become essential. Automated age and gender classification in facial image analysis is a primary process that finds useful in several real time applications such as target advertisements, forensics, human computer, etc. But age and gender classification remains a challenging process due to differences in visual angle, facial expression, background, and facial image appearance. It is more challenging in the un-constrained imaging conditions. In this aspect, this article introduces an automated deep convolutional neural network based age and gender classification (ADCNN-AGC) model using facial images. The proposed ADCNN-AGC model aims for determining the age and gender of the persons who exist in the facial images. To accomplish this, the ADCNN-AGC model follows a two stage process namely face recognition and age/gender classification. Primarily, Multi-task Cascaded Convolutional Networks (MTCNN) model is utilized for the detection of faces in the input images. Besides, the Efficient Net model is applied for the proper extraction of feature vectors which are then passed into the One-Dimensional Convolutional Neural Network (1-DCNN) for classification procedure. The performance validation of the ADCNN-AGC model has been tested using benchmark datasets and the outcomes are observed in many aspects. The experimental outcomes reported the enhanced performance of the ADCNN-AGC model over recent state of art approaches.

Article Details