Facial Expression Recognition using SVM with CNN and Handcrafted Features
G. Priyanka1, S. Pavithra2
1G. Priyanka*, Assistant Professor (Senior Grade), Department of CSE, Mepco Schlenk Engineering College, Sivakasi, India.
2S. Pavithra, PG Student, Department of CSE, Mepco Schlenk Engineering College, Sivakasi, India.
Manuscript received on November 20, 2019. | Revised Manuscript received on November 26, 2019. | Manuscript published on 30 November, 2019. | PP: 3570-3575 | Volume-8 Issue-4, November 2019. | Retrieval Number: D7802118419/2019©BEIESP | DOI: 10.35940/ijrte.D7802.118419
Open Access | Ethics and 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: The facial expression recognition system is playing vital role in many organizations, institutes, shopping malls to know about their stakeholders’ need and mind set. It comes under the broad category of computer vision. Facial expression can easily explain the true intention of a person without any kind of conversation. The main objective of this work is to improve the performance of facial expression recognition in the benchmark datasets like CK+, JAFFE. In order to achieve the needed accuracy metrics, the convolution neural network was constructed to extract the facial expression features automatically and combined with the handcrafted features extracted using Histogram of Gradients (HoG) and Local Binary Pattern (LBP) methods. Linear Support Vector Machine (SVM) is built to predict the emotions using the combined features. The proposed method produces promising results as compared to the recent work in .This is mainly needed in the working environment, shopping malls and other public places to effectively understand the likeliness of the stakeholders at that moment.
Keywords: SVM, CNN, Handcrafted Features, Combined Features, HoG, LBP.
Scope of the Article: Plant Cyber Security.