Recently, one paper on face recognition is
accepted by the IJCV. The full name of IJCV is International Journal of
Computer Vision, which is one of the top journal on computer vision. Below is a
brief introduction of this work.
Chunrui Han, Shiguang Shan, Meina Kan, Shuzhe
Wu, and Xilin Chen, "Personalized Convolution for Face Recognition,"
International Journal of Computer Vision (IJCV), Sep. 2021. (Accepted)
In conventional convolution, the convolutional
kernels for feature extraction are fixed regardless of the input face once the
training stage is finished. By contrast, we humans are usually impressed by some
unique characteristics of different persons, such as one's blue eyes while
another one's naevus at specific location. Inspired by this observation, we
propose a personalized convolution method which aims to extract special
distinguishing characteristics of each person for more accurate face
recognition. Specifically, given a face, we adaptively generate a set of
kernels for him/her, named by us ordinary kernel, which is further analytically
decomposed into two orthogonal components, i.e., the commonality component and
the specialty component. The former characterizes the commonality among
subjects which is optimized on a reference set. The latter is the residual part
by filtering out the commonality component from the ordinary kernel, so as to
capture those special characteristics, named by us personalized kernel. The
CNNs with personalized kernels for convolution can highlight those specialty of
a person's distinguishing characteristics, leading to better distinguishing of
different faces. Additionally, as a by-product, the reference set also
facilitates the adaptation of our method to different scenarios by simply
selecting faces of a particular population. The framework of our method is
below.
Overall, the main contributions of our work lie
in three aspects: 1) we propose the personalized convolution to adaptively
extract the special distinguishing features of an input face for adaptive and
better face recognition; 2) a general framework decomposing one's ordinary
kernel into commonality component and specialty component is proposed to
generate the personalized kernel for extracting one's special distinguishing
features. A reference set is introduced and serves as a strainer to filter out
the commonality component from the ordinary kernel while leaving the specialty
component as the personalized kernel. Different constitutions of the reference
set derive distinct variants of this framework with potential for varying
application scenarios; 3) extensive experiments demonstrate the effectiveness
of personalized feature extraction for face recognition.
Download: