Identifying people’s identity from a group photo through face
recognition models has applications in various fields. There
are two major challenges, first due to the presence of several
faces with various degrees of clarity and scale, and second due
to angular orientation of faces in usual group photos. Detect-
ing and cropping the faces have been reasonably solved using
various segmentation-like models. Recognizing identity after
cropping a frontal face has also been successful to some ex-
tent. However, the presence of orientation, often manifested
by yaw and pitch reduces identifying features from faces. In
such cases, models often make mistakes if they are forced to
match a face with some identity. Our objective is to modulate
the loss based on the orientation angle and consider the fact
into the model that it is harder to detect if the angle of orien-
tation is more. This approach leads to better accuracy and is
also more intuitive. None of the existing methods in the liter-
ature address this problem, and when we compare with them
we found the proposed model to outperform them on several
datasets.