Age and Gender Estimation from Face Images Using CNN-Based Wide Residual Networks Gupta Varun1,*, Jindal Shubham2,** 1Associate Professor, Department of Computer Science and Engineering, Chandigarh College of Engineering and Technology, Chandigarh, India 2Student, Department of Computer Science and Engineering, Chandigarh College of Engineering and Technology, Chandigarh, India *(Corresponding author) email id: varun3dec@gmail.com
**shubham27jindal@gmail.com
Abstract Age and gender are one of the chief personal traits of humans. The estimation of human age and gender is very important to a rising amount of utilisations, especially after the boom of social media and social networking platforms. The popularity of this field is growing day by day due to its vast emerging real world applications, such as forensic art, electronic customer relationship management, law enforcement, security control and surveillance monitoring, biometrics, entertainment, cosmetology, human–computer interactions and many more. In this paper, a deep learning-based solution is being proposed to implement the task of age and gender detection of humans from facial images. For this purpose, facial landmarks of the human faces are not used in the proposed solution. The proposed model has been trained using IMDB–WIKI data set which is the largest public data set for age and gender prediction having 0.5 million facial images from the IMDB and Wikipedia web domains. In the proposed work, the IMDB data set has been used as the training data set, whereas WIKI has served as the test data set. The backbone of the proposed system consists of Convolutional Neural Network-based Wide Residual Networks. These networks are computationally inexpensive and also provide state-of-the-art results on several competitive benchmarks. The results indicate that the proposed system performs better than the existing systems for age and gender estimation task. Top Keywords Age and gender estimation, Convolutional Neural Network (CNN or ConvNet), Wide Residual Network (WideResNet), Neural networks, Deep learning, Machine learning, Artificial intelligence. Top |