Main Article Content
Identification of human soft biometrics is a study field on hot computer vision (CV). This is mainly due to the greater dependency on monitoring systems which provide huge quantities of visual data to be analyzed off-line. Smartphones are also a major player where you can buy these soft biometrics with a normal smartphone camera. Most soft biometric identification systems are supported on human facial analysis, in which the human face is an identifiable demographic key feature. So Eyeglasses detection is necessary thing for this. Due to its possible use in a vast number of applications, such as access control, human computer interface and automated search of a large-scale facial picture data base, automatic facial recognition has become an important subject in study. One key need for effective face identification is the resilience of changes in light conditions, facial expressions, postures, scales and other objects' occlusion. Glasses are one of the most frequent occlusive items among occluding objects and have a substantial impact on the performance of facial detection systems.
In face recognition systems, there are two methods to addressing occlusion issues. One of the stages is to estimate the non-occcluded picture of the face. The alternative is to identify and exclusively utilize non-occlusive areas. Hwang and Lee have attempted to rebuild the face picture using the linear square solution of the morphable face model from partly occult faces. However, pixel-specific correlation between input and reference face should be accurately determined or computed according to the morphable model. For example, Martinez has devised a local match method for a strong identification of partly obscured faces by sunglasses and scarves without reconstruction or removal