DeepID3: Face Recognition with Very Deep Neural Networks

By Yi Sun, Ding Liang, Xiaogang Wang, Xiaoou Tang. 2015

975 citations recorded by the source corpus.

The state-of-the-art of face recognition has been significantly advanced by the emergence of deep learning. Very deep neural networks recently achieved great success on general object recognition because of their superb learning capacity. This motivates us to investigate their effectiveness on face recognition. This paper proposes two very deep neural network architectures, referred to as DeepID3, for face recognition. These two architectures are rebuilt from stacked convolution and inception layers proposed in VGG net and GoogLeNet to make them suitable to face recognition. Joint face identification-verification supervisory signals are added to both intermediate and final feature extraction layers during training. An ensemble of the proposed two architectures achieves 99.53% LFW face verification accuracy and 96.0% LFW rank-1 face identification accuracy, respectively. A further discussion of LFW face verification result is given in the end.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

How Image Degradations Affect Deep CNN-Based Face Recognition? (2016, background): “With the recent advances in deep convolutional neural networks, researchers have reached encouraging improvements in face recognition accuracy on the LFW [1] and YouTube Faces [2] datasets [3], [4], [5].”
SphereFace: Deep Hypersphere Embedding for Face Recognition (2017, methodology): “To address this, some methods combine softmax loss with contrastive loss [26, 27, 30] or center loss [37] to enhance the discrimination power of features.”
The MegaFace Benchmark: 1 Million Faces for Recognition at Scale (2015, methodology): “Face recognition has seen major breakthroughs in the last couple of years, with new results by multiple groups [25, 29, 27] surpassing human performance on the leading Labeled Faces in the Wild (LFW) benchmark [15] and achieving near perfect results.”
Sampling Matters in Deep Embedding Learning (2017, background): “DeepID2 [30] and DeepID3 [31] use 25 networks on 25 face regions based on positions of facial landmarks.”
MS-Celeb-1M: A Dataset and Benchmark for Large-Scale Face Recognition (2016, background): “Recently, the interest in the other type of face recognition task, face identiﬁ-cation, has greatly increased [9,10,11,3].”
CosFace: Large Margin Cosine Loss for Deep Face Recognition (2018, background): “Recent works DeepID2+ [21] and DeepID3 [22] further explore the advanced network structures to boost performance.”
Discrimination-aware Channel Pruning for Deep Neural Networks (2018, background): “, image classification [19, 38, 7], face recognition [34, 39], object detection [32, 33], image captioning [44, 47] and video analysis [35, 45].”
One-shot Face Recognition by Promoting Underrepresented Classes (2017, background): “Recently, we have seen a considerable amount of research results re-ported on LFW, which are obtained by using different algorithms on different datasets [1, 17, 16, 18, 15, 19, 20].”
Multimodal Gesture Recognition Using 3-D Convolution and Convolutional LSTM (2017, background): “With the rapid development of the deep learning theory [10], data-driven methods have demonstrated amazing performances in image classification [13], image segmentation [15], object detection [16], scene recognition [18], face recognition [19], human action recognition [20], and human gesture recogn”
When Face Recognition Meets with Deep Learning: An Evaluation of Convolutional Neural Networks for Face Recognition (2015, background): “Finally DeepID3 [22] also modified two famous networks: Inception [26] and VGG Net [20] by adding supervision information to each layer and network ensemble fusion.”
Past, Present, and Future of Face Recognition: A Review (2020, neutral): “[119] 2015 VGGNet 25 Contrastive Softmax + JB WDRef + CelebFaces + (290 k,12 k) 99.”
Shepard Convolutional Neural Networks (2015, background): “In the past a few years, deep learning has been very successful in addressing many aspects of visual perception problems such as image classification, object detection, face recognition [1, 2, 3], to name a few.”
Recent Advances in Deep Learning for Object Detection (2019, background): “In fact, object detection is the basic step towards many computer vision applications, such as face recognition [5, 6, 7], pedestrian detection [8, 9, 10], video analysis [11, 12], and logo detection [13, 14, 15].”
DSOD: Learning Deeply Supervised Object Detectors from Scratch (2017, background): “The effectiveness of deeply supervised learning has been demonstrated in GoogLeNet [32], DSN [18], DeepID3 [30], etc.”
Face Model Compression by Distilling Knowledge from Neurons (2016, methodology): “In particular, the selection is motivated by three original observations (domain knowledge) of face representation disclosed in this work, which are naturally generalized to all DNNs trained by distinguishing massive identities, such as [19, 23, 24, 22].”
Predicting survival from colorectal cancer histology slides using deep learning: A retrospective multicenter study (2019, background): “Applications of CNNsrangewidely,from speechrecognition[14,15],facerecognition[16], or traffic sign classification[17] to masteringtheJapanese gameof Go [18].”
Learning Compatible Embeddings (2021, background): “Visual search or retrieval systems [37, 38] are widely used in many real-world applications such as face recognition [43, 40, 34, 27, 28], person re-identiﬁcation [41, 17, 35], car re-identiﬁcation [20] and image retrieval [2, 12].”
OFF-ApexNet on Micro-expression Recognition System (2018, background): “Deep learning has signiﬁcant impacts on a variety of application domains as it yields numerous state-of-the-art results, such as speech recognition [43], face recognition [44] and scene recognition [45].”
Do We Really Need to Collect Millions of Faces for Effective Face Recognition? (2016, background): “Since then, CNN based recognition systems continuously cross performance barriers with some notable examples including the Deep-ID 1-3 systems [34,32,33].”
A Survey of Convolutional Neural Networks: Analysis, Applications, and Prospects (2020, methodology): “It was first used in SiameseNet [67] and later was deployed in DeepID2 [68], DeepID2+ [69], and DeepID3 [70].”