Cross-Batch Hard Example Mining With Pseudo Large Batch for ID vs. Spot Face Recognition

doi:10.1109/TIP.2021.3137005

CORC > 自动化研究所 > 中国科学院自动化研究所 > 模式识别国家重点实验室 > 生物识别与安全技术研究中心

	Cross-Batch Hard Example Mining With Pseudo Large Batch for ID vs. Spot Face Recognition
	Tan, Zichang 7,8; Liu, Ajian 6; Wan, Jun 3,4,5; Liu, Hao 3,4; Lei, Zhen 2,3,4; Guo, Guodong 7,8; Li, Stan Z.1,5
刊名	IEEE TRANSACTIONS ON IMAGE PROCESSING
	2022
卷号	31 页码:3224-3235
关键词	Face recognition Training Measurement Graphics processing units Deep learning Logic gates Feature extraction Face recognition ID vs spot deep learning cross-batch hard example mining pseudo large batch
ISSN号	1057-7149
DOI	10.1109/TIP.2021.3137005
通讯作者	Wan, Jun(jun.wan@nlpr.ia.ac.cn)
英文摘要	In our daily life, a large number of activities require identity verification, e.g., ePassport gates. Most of those verification systems recognize who you are by matching the ID document photo (ID face) to your live face image (spot face). The ID vs. Spot (IvS) face recognition is different from general face recognition where each dataset usually contains a small number of subjects and sufficient images for each subject. In IvS face recognition, the datasets usually contain massive class numbers (million or more) while each class only has two image samples (one ID face and one spot face), which makes it very challenging to train an effective model (e.g., excessive demand on GPU memory if conducting the classification on such massive classes, hardly capture the effective features for bisample data of each identity, etc.). To avoid the excessive demand on GPU memory, a two-stage training method is developed, where we first train the model on the dataset in general face recognition (e.g., MS-Celeb-1M) and then employ the metric learning losses (e.g., triplet and quadruplet losses) to learn the features on IvS data with million classes. To extract more effective features for IvS face recognition, we propose two novel algorithms to enhance the network by selecting harder samples for training. Firstly, a Cross-Batch Hard Example Mining (CB-HEM) is proposed to select the hard triplets from not only the current mini-batch but also past dozens of mini-batches (for convenience, we use batch to denote a mini-batch in the following), which can significantly expand the space of sample selection. Secondly, a Pseudo Large Batch (PLB) is proposed to virtually increase the batch size with a fixed GPU memory. The proposed PLB and CB-HEM can be employed simultaneously to train the network, which dramatically expands the selecting space by hundreds of times, where the very hard sample pairs especially the hard negative pairs can be selected for training to enhance the discriminative capability. Extensive comparative evaluations conducted on multiple IvS benchmarks demonstrate the effectiveness of the proposed method.
资助项目	National Key Research and Development Plan[2021YFE0205700] ; External Cooperation Key Project of Chinese Academy Sciences[173211KYSB20200002] ; Chinese National Natural Science Foundation[62106264] ; Chinese National Natural Science Foundation[61876179] ; Chinese National Natural Science Foundation[61961160704] ; Key Project of the General Logistics Department[AWS17J001] ; Science and Technology Development Fund of Macau[0008/2019/A1] ; Science and Technology Development Fund of Macau[0025/2019/AKP] ; Science and Technology Development Fund of Macau[0070/2020/AMJ]
WOS关键词	MODEL
WOS研究方向	Computer Science ; Engineering
语种	英语
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号	WOS:000794186400003
资助机构	National Key Research and Development Plan ; External Cooperation Key Project of Chinese Academy Sciences ; Chinese National Natural Science Foundation ; Key Project of the General Logistics Department ; Science and Technology Development Fund of Macau
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/49404]
专题	自动化研究所_模式识别国家重点实验室_生物识别与安全技术研究中心
通讯作者	Wan, Jun
作者单位	1.Westlake Univ, Sch Engn, Hangzhou 310024, Peoples R China 2.Chinese Acad Sci, Hong Kong Inst Sci & Innovat, Ctr Artificial Intelligence & Robot, Hong Kong, Peoples R China 3.Univ Chinese Acad Sci UCAS, Sch Artificial Intelligence, Beijing 100190, Peoples R China 4.Chinese Acad Sci CASIA, Inst Automat, Natl Lab Pattern Recognit NLPR, Beijing 100190, Peoples R China 5.Macau Univ Sci & Technol MUST, Fac Innovat Engn, Taipa, Macao, Peoples R China 6.Macau Univ Sci & Technol MUST, Fac Innovat Engn, Macau, Peoples R China 7.Natl Engn Lab Deep Learning Technol & Applicat, Beijing 100000, Peoples R China 8.Baidu Res, Inst Deep Learning, Beijing 100000, Peoples R China
推荐引用方式 GB/T 7714	Tan, Zichang,Liu, Ajian,Wan, Jun,et al. Cross-Batch Hard Example Mining With Pseudo Large Batch for ID vs. Spot Face Recognition[J]. IEEE TRANSACTIONS ON IMAGE PROCESSING,2022,31:3224-3235.
APA	Tan, Zichang.,Liu, Ajian.,Wan, Jun.,Liu, Hao.,Lei, Zhen.,...&Li, Stan Z..(2022).Cross-Batch Hard Example Mining With Pseudo Large Batch for ID vs. Spot Face Recognition.IEEE TRANSACTIONS ON IMAGE PROCESSING,31,3224-3235.
MLA	Tan, Zichang,et al."Cross-Batch Hard Example Mining With Pseudo Large Batch for ID vs. Spot Face Recognition".IEEE TRANSACTIONS ON IMAGE PROCESSING 31(2022):3224-3235.