Abstract :
— In many real-world scenarios especially some
unconstrained environments, human faces are occluded by
other objects, therefore it is difficult to obtain fully holistic
face images for recognition. Most ancient face recognition
algorithms have faith in face alignment and face
standardization, which is highly infeasible. To address this,
a effective partial face recognition approach is required to
recognize persons of interest from their partial faces A new
partial face recognition approach based on Directional
Binary Code is proposed. It is used for automatic
conversion of partial faces to gallery faces by aligning
partial face patches to gallery faces automatically. DBC is
applied on image to encode the directional edge information.
It captures the spatial relationship between any pair of
neighborhood pixels in a local region along a given
direction. The extracted features from the given image are
used for face recognition. It reflects the image local feature.
It is robust to occlusions as well as illumination changes.
Keyword :
Face recognition, partial face recognition, feature set matching, feature alignment, image matching, bio metrics.