论文速看[2019-1-5]-Face Recognition: A Novel Multi-Level Taxonomy based Survey

一篇关于人脸识别中多任务的综述，submit to IET Biometrics
https://arxiv.org/ftp/arxiv/papers/1901/1901.00713.pdf

摘要：

In a world where security issues have been gaining growing importance, face recognition systems have attracted increasing attention in multiple application areas, ranging from forensics and surveillance to commerce and entertainment. To help understanding the landscape and abstraction levels relevant for face recognition systems, face recognition taxonomies allow a deeper dissection and comparison of the existing solutions. This paper proposes a new, more encompassing and richer multi-level face recognition taxonomy, facilitating the organization and categorization of available and emerging face recognition solutions; this taxonomy may also guide researchers in the development of more efficient face recognition solutions. The proposed multi-level taxonomy considers levels related to the face structure, feature support and feature extraction approach. Following the proposed taxonomy, a comprehensive survey of representative face recognition solutions is presented. The paper concludes with a discussion on current algorithmic and application related challenges which may define future research directions for face recognition.

文章亮点

1. The proposed taxonomy considers face structure and feature support levels for a more complete characterization of face recognition solutions. These levels provide critical information on the way face recognition solutions deal with the facial structure and what is the region of spatial support for feature extraction, impacting the selection of the corresponding approach.
   强调face structure和feature support在人脸识别算法中的作用。
2. The feature extraction approach level in the proposed taxonomy (equivalent to the matching features level in [10]) considers a richer and more complete set of classes, rather than only feature based and appearance based classes as in [10].
3. The feature extraction approach level is further divided into feature extraction sub-approaches to more deeply understand the technological landscape in the context of feature extraction.
   本文涉及的feature extraction比文献[10]更加复杂，考虑的更多，并且进行更深层次的划分

Proposing a novel multi-level face recognition taxonomy

本文的分类结构如上图所示，包括4个部分

1. Face structure：如何处理人脸，3种方法
   (i) global representation全局表征，将人脸看作一个整体;
   (ii) component + structure representation部分+结构表征，将人脸看作由各个有关系的子部分组成;
   (iii) component representation部分表征，将人脸看作由各个无关的部分组成。
2. Feature support：特征支撑面，分为全局特征和局部特征
3. Feature extraction approach:特征提取方法
   (i)appearance based,deriving features by using statistical transformations from the intensity data;
   (ii)model based,deriving features based on geometrical characteristics of the face;
   (iii)learning based,deriving features by modelling and learning relationships from the input data;
   (iv)hand-crafted based,deriving features from pre-selected elementary characteristics.
4. Feature extraction sub-approach对上面的4种特征提取方法更进一步分析

Face recognition: status quo

作者做了一个2页的table2对现有的人脸识别方法进行了分析，对提出的4个层面进行了总结。

1. Face structure：80%的文章采用global representation，15%的文章采用 component + structure representation，只有~5%采用 component representation；
2. Feature support：global和local选择的人一样多，主要取决于特征提取方法；
3. Feature extraction approach: learning和hand-crafted较多，且随着deep learning的发展，learning based的方法逐渐成为了主流；
4. Feature extraction sub-approach：超过80%的方法是learning based

Challenges and future research directions

最后，从算法和应用两个方面，作者对现在人脸识别存在的问题与挑战，和以后的研究方向进行了阐述，列出了以下发展趋势。

Algorithmic research challenges

• Double-deep face recognition for sequential biometric data
• Face recognition for new sensors
• Deep face recognition model compression
• Face presentation attack detection for face recognition security
• Face identification from photo-sketches

Application related challenges

• Face recognition in embedded systems
• Face authentication in mobile phones
• Face recognition in autonomous vehicles
• Face analysis in health informatics
• Face recognition in smart cities