Human Emotion Recognition Using Mean of Average and Maximum Pooling
C. Karthik1, D. Chandrasekhar2, B. Naveen Kumar3, B. Naveen4
1C. Karthik, Associate Professor, Department of ECE, QIS Group Colleges, Ongole (Andhra Pradesh), India.
2D. Chandrasekhar, Assistant Professor, Department of ECE, Siddhartha Institute of Technology & Sciences, Narapally, Ghatkesar, Hyderabad (Telangana), India.
3B. Naveen Kumar, Assistant Professor, Department of ECE, Siddhartha Institute of Technology & Sciences, Narapally, Ghatkesar, Hyderabad (Telangana), India.
4B. Naveen, Assistant Professor, Department of ECE, Siddhartha Institute of Technology & Sciences, Narapally, Ghatkesar, Hyderabad (Telangana), India.
Manuscript received on 08 December 2019 | Revised Manuscript received on 22 December 2019 | Manuscript Published on 31 December 2019 | PP: 365-368 | Volume-8 Issue-12S2 October 2019 | Retrieval Number: L107010812S219/2019©BEIESP | DOI: 10.35940/ijitee.L1070.10812S219
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open-access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Currently a the very beginning’s of the unsolved difficulty in pc imaginative and prescient is perceiving or understanding different people’ feelings and sentiments. Albeit ongoing strategies accomplish close to human exactness in controlled conditions, the acknowledgment of emotions inside the wild remains a hard difficulty. On this paper we proposed MAM Pooling (mean of common and maximum) strategy with CNN to perceive human feelings. We center round programmed distinguishing evidence of six emotions constantly: Happiness, Anger, unhappiness, surprise, fear, and Disgust. Convolutional Neural network (CNN) is a certainly propelled trainable layout that may study invariant highlights for numerous programs. Whilst all is said in carried out, CNNs include of rotating convolutional layers, non-linearity layers and highlight pooling layers. In this artwork, a Novel include pooling approach, named as MAM pooling is proposed to regularize CNNs, which replaces the deterministic pooling obligations with a stochastic system thru taking the ordinary of max pooling and regular pooling strategies. The benefit of the proposed MAM pooling technique lies in its first-rate capability to address the over fitting problem skilled with the resource of CNN age.
Keywords: Emotion; Face Expression; MAM Pooling, CNN.
Scope of the Article: Pattern Recognition