Symmetry · 2019 · 683 citations · 94 references
ComputingComputer ScienceData MiningMachine LearningStatistical LearningDeep LearningKernel MethodsMetric LearningNeural NetworksRepresentation LearningActivation FunctionsDeep Metric LearningSiamese NetworksTriplet Networks
Metric learning seeks to quantify sample similarity with an optimal distance metric, but linear projections struggle with nonlinear data; kernel methods and, more recently, deep metric learning with activation functions have been adopted to address this, with Siamese and Triplet networks demonstrating success by capturing similarity relationships. The article aims to highlight the significance of deep metric learning and the challenges addressed in recent research. The review focuses on Siamese‑ and Triplet‑inspired deep metric learning models that share weights, emphasizing how sampling strategies, distance metrics, and network architecture influence performance. It is the first comprehensive survey that systematically analyzes and evaluates these factors, comparing quantitative results across methods.
Metric learning aims to measure the similarity among samples while using an optimal distance metric for learning tasks. Metric learning methods, which generally use a linear projection, are limited in solving real-world problems demonstrating non-linear characteristics. Kernel approaches are utilized in metric learning to address this problem. In recent years, deep metric learning, which provides a better solution for nonlinear data through activation functions, has attracted researchers’ attention in many different areas. This article aims to reveal the importance of deep metric learning and the problems dealt with in this field in the light of recent studies. As far as the research conducted in this field are concerned, most existing studies that are inspired by Siamese and Triplet networks are commonly used to correlate among samples while using shared weights in deep metric learning. The success of these networks is based on their capacity to understand the similarity relationship among samples. Moreover, sampling strategy, appropriate distance metric, and the structure of the network are the challenging factors for researchers to improve the performance of the network model. This article is considered to be important, as it is the first comprehensive study in which these factors are systematically analyzed and evaluated as a whole and supported by comparing the quantitative results of the methods.
94
Gradient-based learning applied to document recognition
Yann LeCun, Léon Bottou, Yoshua Bengio et al. · Proceedings of the IEEE · 1998 · 58.2K citations · Full text
FaceNet: A unified embedding for face recognition and clustering
2015 · 11.2K citations · Full text
Dimensionality Reduction by Learning an Invariant Mapping
Raia Hadsell, Sumit Chopra, Yann LeCun · 2006 · 5.2K citations · Full text
Prototypical Networks for Few-shot Learning
Jake Snell, Kevin Swersky, Richard S. Zemel · arXiv (Cornell University) · 2017 · 5.2K citations · Full text
Scalable Person Re-identification: A Benchmark
Liang Zheng, Liyue Shen, Lu Tian et al. · 2015 · 4.6K citations