Robust Face Recognition Under Pose, Illumination and Expression Variations using L1 Graph method
Several learning techniques have been used to train the classifiers for face recognition, such as SVM. Although applying an appropriate dimension reduction algorithm or a robust classification technique may yield more accurate recognition results, they usually require multiple training images for each subject. However, multiple training images per subject may not be available in practice. Also the systems were focused on single solution which cannot deal with multiple issues combined together such as Pose Variation, Illumination Variation and Expression Variant human faces. Additionally those systems were able to provide solution for monochrome images, where color images are considered as complex issue.
The Major flaw of face recognition system are due to the chromatic variances , which increases the complexity of recognizing an individual’s face. The illumination changes are highly discussed in many researches as a primary drawback of face recognition, whenever the light conditions changes in capturing the facial information, the results also unexpected which makes the system more unpredictable. The researchers agree the fact of grayscale faces are producing better results when compared to Multicolor 2D RGB Faces.
The next issue that makes even human to get illusion is in understanding the pose of human face, when ever the alignment of face and angle of capturing changes, the accuracy of recognizing the exact human remains unanswered . the complexity of approximating the individual get additional degradation to the previously discussed illumination variances.
Various Human- computer Interfacing systems and robotic vision based system works on understanding the timely expression of human faces , which are under research for almost a decade. Human to human expression classification itself a difficult task,...