Paper Reading on Face Recognition

2019.03.28
ludics

1. ArcFace: Additive Angular Margin Loss for Deep Face Recognition

Abstract

Design of loss function that enhance discriminative power.

ArcFace: Additive Angular Margin Loss, has a clear geometric interpretation due to the geodesic distance on the hypersphere.

Introduction

DCNNs map face image into a feature that has small intra-class and large inter-class distance.

Two lines to train:

  1. multi-class classifier, separates different identities using softmax
  2. learn an embedding, triplet loss(anchor, positive, negative)

==> Both have drawbacks

WRd×n W \in \mathbb{R}^{d \times n}
  1. Softmax loss: increases linearly with n; learned feature separable for closed-set classification but not discriminative
  2. Triplet loss: combinatorial explosion; semi-hard sample mining difficult

Enhanced softmax:

  1. Centre loss by Wen et al., obtain intra-class compactness; updating centres is difficult.
  2. multiplicative angular margin penalty, enforce intra-class compactness and inter-class discrepancy, leading to better discriminative. Angular margin, Sphereface.
  3. CosFace, adds cosine margin penalty to the target logit

This paper: ArcFace, further improve discrimanative power of face recognition & stabilise training process. Dot product between feature & last fc layer equal to the cosine distance after feature & weight norm. Use arc-cosine to cal the angle between current feature & target weight. Add an additive angular margin to target angle, get target logit back by cosine. Re-scale logits by fixed feature norm, subsequent same as softmax.

Advantages:

Proposed Approach

ArcFace

Softmax loss:

L1=1Ni=1NlogeWy(i)Txi+by(i)j=1neWjTxi+bj L_1 = -\frac{1}{N} \sum_{i=1}^N \log \frac{e^{W_{y^{(i)}}^T x_i + b_{y^{(i)}}}}{\sum_{j=1}^n e^{W_j^T x_i + b_j}}

2. Support Vector Guided Softmax Loss for Face Recognition

SV-Softmax, inherit the advantages of mining-based and margin-based losses into one framework.

3. Large-scale Bisample Learning on ID versus Spot Face Recognition

Mainly two new methods:

4. Orthogonal Deep Features Decomposition for Age-Invariant Face Recognition

Main Formulas:

x=xagexid x = x_{age} \cdot x_{id}
xage=x2 x_{age} = \parallel x \parallel _2
xid={x1x2,x2x2,,xnx2} x_{id} = \lbrace \frac{x_1}{\parallel x \parallel _2}, \frac{x_2}{\parallel x \parallel _2}, \cdots, \frac{x_n}{\parallel x \parallel _2} \rbrace

Main purpose: Look the L2 length of represetation as age

5. Quality Aware Network for Set to Set Recognition

By Liu Yu

Main Aim: Get the quality of image as weight to build a set represetation

6. Linkage Based Face Clustering via Graph Convolution Network

by Wang Zhongdao

GCN method to cluster face images

7. Learning Face Representation from Scratch

CASIA WebFace

Collect From IMDB,

8. IARPA Janus Benchmark-B Face Dataset

IJB Dataset