Discriminative Shared Gaussian Process Based On Latent Variable Model – An Approach for Facial Expression Recognition
Journal Title: International Journal for Research in Applied Science and Engineering Technology (IJRASET) - Year 2016, Vol 4, Issue 3
Abstract
Facial expression is an important channel for human communication and can be applied in many real applications. One critical step for facial expression recognition (FER) is to accurately extract emotional features. Existing methods for multiview and/or view-invariant facial expression recognition typically perform separation of the observed expression using either classifiers learned separately for each view or a single classifier learned for all views. However, these approaches ignore the fact that different views of a facial expression are just different manifestations of the same facial expression. By accounting for this idleness, we can design more efficient classifiers for the target task. This paper proposes a discriminative shared Gaussian process latent variable model (DS-GPLVM) for multi-view and view- invariant separation of facial expressions from multiple views [1]. In this model, first learn a discriminative manifold shared by multiple views of a facial expression. Subsequently, we perform facial expression separation in the expression manifold. Finally, separation of an observed facial expression is carried out either in the view-invariant manner (using only a single view of the expression) or in the multi-view manner (using multiple views of the expression). The proposed model can also be used to perform mixture of different facial features in a principled manner. DS-GPLVM is proposed on both posed and impulsively displayed facial expression from three publicly available datasets (Multi-PIE, labeled face parts in the wild and static facial expression in the wild). And the results show that this model outperforms the modern methods for multi-view and view-invariant facial expression separation, and several modern methods for multi-view learning and feature mixture.
Authors and Affiliations
R. Keerthanadevi, R. Sureshkumar
Keywords
Related Articles
- EP ID EP21736
- DOI -
- Views 254
- Downloads 4
Study of Efficient Firewall Packet Filtering and Matching Algorithms
for the network devices like firewall or IPsec, packet filtering is plays extremely significant role in high speed networks. Thus it is important that firewall policies should be optimized in order to make available the...
Survey on Speckle Filtering Techniques of the Ultrasound Images
Today, ultrasound and magnetic resonance images are the essential tools for non-invasive medical diagnosis. One of the fundamental problems in this area is speckle noise, which is a major limitation on the image quality...
IR Sensor Based Squats Detection of Railway Track using Robot
This paper proposes the design of crack finding robot for finding cracks in the railway tracks. Here the LPC2148 microcontroller is interfaced with Robot, Global Positioning System (GPS), Liquid Crystal Display (LCD) an...
Parameter Optimization of Gas Metal Arc Welding Using Taguchi Method
Gas metal arc welding (GMAW) is currently one of the most popular welding methods, especially in industrial environments. In order to meet the global competition and the survival of products in the market a new way of t...
Site Locating and Insite Exploring Smart Crawers For Deep Web Interfaces
Mobile Android applications typically have access to sensitive knowledge and resources on the user device. Misuse of this knowledge by malicious applications could result in privacy breaches and sensitive data outpourin...