A Modular Approach For Facial Expression Recognition using Euler Principal Component Analysis (e-PCA)
Sujata, M.M. Trivedi, Suman Kumar Mitra · 2018
Though spectacular recognition rates are achieved with numerous techniques underneath the controlled face image capturing surroundings, creating recognition a lot of reliable underneath uncontrolled surroundings remains an excellent challenge. Security and surveillance images, pictures, captured in open uncontrolled environments, ar seemingly subjected to extreme lighting conditions like underexposed, and overexposed areas that scale back the number of helpful details obtainable within the collected face images. Euler Principal Component Analysis uses a difference live to extend the variations between objects although the face images area unit below the influence of visual variation. In this paper, we investigate e-PCA that can be used as the facial expression recognition in modular way. 1-NN classifier is used for our recognition purpose. Experiments area unit done on JAFFE, VIDEO and CK+ databases and it shown that there area unit enhancements within the expression recognition rate using Euler Principal Component Analysis (e-PCA) under certain circumstances.