Cosine Similarity Metric Learning for Face Verification

By Hieu V. Nguyen, Li Bai. 2010

766 citations recorded by the source corpus.

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

FairNAS: Rethinking Evaluation Fairness of Weight Sharing Neural Architecture Search (2019, background): “But how much do they resemble each other? We hereby involve the cosine similarity [31] to measure the distance among various feature vectors.”
Hybrid Deep Learning for Face Verification (2013, methodology): “At the recognition stage, classifiers such as SVM are used to classify two face images as having the same identity or not [5, 24, 13], or other models are employed to compute the similarities of two face images [10, 22, 12, 6, 7, 25].”
Learning Compact Binary Face Descriptor for Face Recognition (2015, unknown): “CSML+SVM, aligned+WPCA [43] 88:00 0:37 PAF [67] 87:77 0:51 SFRD+PMML+WPCA [11] 89:35 0:50 Sub-SML [7] 89:73 0:38 VMRS+WPCA [3] 91:10 0:59 DDML+WPCA [19] 90:68 1:41 CBFD+WPCA(a) 87:33 2:42 CBFD+WPCA(b) 87:57 1:43 CBFD+WPCA(c) 87:23 1:68 CBFD+WPCA(mean: a, b, c) 89:05 1:51 CBFD+WPCA(svm: a, b, c) 89:0”
Deep Metric Learning for Practical Person Re-Identification (2014, methodology): “Because of the good property of Cosine function and it has been used widely in many pattern recognition problems [3], [23], [24], we choose it as the connection function.”
Discriminative Deep Metric Learning for Face Verification in the Wild (2014, methodology): “As suggested in [16, 26, 36], we also use the square root of each feature and evaluate the performance of our DDML method when all the six different feature descriptors are combined.”
A survey on instance selection for active learning (2012, methodology): “Nguyen and Li [38] proposed a cosine similarity metric for face verification.”
Implicit Diversity in Image Summarization (2019, background): “While cosine distance is considered appropriate for measuring similarity between images [65], in general, identifying the appropriate metric to quantify pairwise similarity between two individuals is another active topic of research and we refer the readers to [61, 90] for further discussion in this”
LightFace: A Hybrid Deep Face Recognition Framework (2020, methodology): “Herein, we can calculate cosine similarity [29], Euclidean distance [30] or its L2 form to find distances between vectors.”
ArcFace: Additive Angular Margin Loss for Deep Face Recognition (2018, unknown): “L2-normalised Euclidean distance and cosine distance [29].”
Fisher Vector Faces in the Wild (2013, background): “The question that this paper addresses is what happens if, instead of developing yet another face-specific image representation, one applies off-the-shelf object recognition representations to AFV.”
PCCA: A new approach for distance learning from sparse pairwise constraints (2012, unknown): “[22] reports 85.”
Learning hierarchical representations for face verification with convolutional deep belief networks (2012, methodology): “In practice, we find that using ITML improves performance over CSML by several percentage points.”
Learning to Align from Scratch (2012, methodology): “(4)We note that the accuracy published in [40] was higher than we were able to obtain in our own implementation.”
Neighborhood repulsed metric learning for kinship verification (2012, methodology): “Representative metric learning algorithms include neighborhood component analysis (NCA) [8], cosine similarity metric learning (CSML) [18], large margin nearest neighbor (LMNN) [23], and information theoretic metric learning (ITML) [4].”
Motion Interchange Patterns for Action Recognition in Unconstrained Videos (2012, methodology): “Then, using the training data, we learn eight transformations Tα by employing the CSML algorithm [28].”