Real time facial expression and gender recognition with feature integrated CNN

ABSTRACT

For social communication, facial expression recognition is crucial; further improvement with additional facial attributes like gender can boost the performance of intelligent applications. Recently, the incorporation of deep learning methods extended the research of facial expressions and gender recognition. The deep learning method gives a better detection rate but incurs high computation costs and memory requirements. In this regard, a feature-integrated convolution neural network, namely FusedEGNET, has been designed with fewer parameters to deploy in real-world environments. Further, it utilizes joint supervision with softmax and centre loss to learn the discriminative features of facial expressions. Besides, evaluations of five publicly available facial expression datasets, namely FER-2013, JAFFE, CK+, KDEF, and FERPlus, along with the Adience and CK + datasets for gender recognition, have also been carried out. Finally, to examine the efficiency of the proposed method, an application has been developed and applied to detect real-time facial expressions and gender.

KEYWORD:

• Convolutional neural networks
• facial expression recognition
• gender recognition
• center loss
• transfer learning

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Tanusree Podder

Tanusree Podder is currently working as a Principal ITI Kamalpur, Tripura, India. She is currently pursuing the Ph.D. degree in Computer Science & Engineering from National Institute of Technology Agartala, Tripura, India. Her research interests include Digital Image Processing, Computer Vision, Deep Learning and Pattern Recognition.

Diptendu Bhattacharya

Dr. Diptendu Bhattacharya received M.E.Tel.E. and Ph.D. (Engineering) from Jadavpur University, Kolkata, India, in 1999 and 2016 respectively. He is currently working as an Associate Professor in the Department of Computer Science and Engineering, National Institute of Technology Agartala, Tripura, India. His research interests include Digital Image Processing, Machine Intelligence in Economic Time series prediction, artificial intelligence, Fuzzy time series, and its prediction. He is a member of the IEEE and the IEEE Computer Society.

Abhishek Majumdar

Dr. Abhishek Majumdar is currently working as an Assistant Professor and Head of the Department of Computer Science and Engineering (Specialization in Artificial Intelligence), Techno India University, Kolkata, India. Dr. Majumdar did his Ph.D. in Machine Learning from the National Institute of Technology Silchar, India. His area of interest includes Deep Learning, Bioinformatics, and Multi objective Optimization. Dr. Majumdar also holds the position of institute coordinator of the Computer Society of India.