Harish H. Kenchannavar

Work place: Department of Information Science and Engineering, Gogte Institute of Technology, Belagavi, Visvesvaraya Technological University, Belagavi, 590018, India

E-mail: harishhk@git.edu


Research Interests:


Harish H. Kenchannavar is working as Professor in the Department of Information Science and Engineering, Gogte Institute of Technology, Belagavi. He has 21 years of teaching experience and 10 years of research experience. His research contributions are in the field of Wireless sensor network, Quality of Service (QoS), Computer Vision and system modeling. He has vast experience in Academic, research and coordination at University, College and Departmental level. He is Life member to computer society of India and ISTE professional bodies. He has published 20 + research papers at national and international conference in India as well as outside country. He was college-level National Accreditation Board (NBA) coordinator. 

Author Articles
Hybrid Deep Optimal Network for Recognizing Emotions Using Facial Expressions at Real Time

By Rakshith M. D. Harish H. Kenchannavar

DOI: https://doi.org/10.5815/ijisa.2024.03.04, Pub. Date: 8 Jun. 2024

Recognition of emotions by utilizing facial expressions is the progression of determining the various human facial emotions to infer the mental condition of the person. This recognition structure has been employed in several fields but more commonly applied in medical arena to determine psychological health problems. In this research work, a new hybrid model is projected using deep learning to recognize and classify facial expressions into seven emotions. Primarily, the facial image data is obtained from the datasets and subjected to pre-processing using adaptive median filter (AMF). Then, the features are extracted and facial emotions are classified through the improved VGG16+Aquila_BiLSTM (iVABL) deep optimal network. The proposed iVABL model provides accuracy of 95.63%, 96.61% and 95.58% on KDEF, JAFFE and Facial Expression Research Group 2D Database (FERG-DB) which is higher when compared to DCNN, DBN, Inception-V3, R-152 and Convolutional Bi-LSTM models. The iVABL model also takes less time to recognize the emotion from the facial image compared to the existing models.

[...] Read more.
Other Articles