Abstract:The objective of deep metric learning (DML) is to learn embeddings that can capture semantic similarity and dissimilarity information among data points. Existing pairwise or tripletwise loss functions used in DML are known to suffer from slow convergence due to a large proportion of trivial pairs or triplets as the model improves. To improve this, ranking-motivated structured losses are proposed recently to incorporate multiple examples and exploit the structured information among them. They converge faster and achieve state-of-the-art performance. In this work, we unveil two limitations of existing ranking-motivated structured losses and propose a novel ranked list loss to solve both of them. First, given a query, only a fraction of data points is incorporated to build the similarity structure. Consequently, some useful examples are ignored and the structure is less informative. To address this, we propose to build a set-based similarity structure by exploiting all instances in the gallery. The learning setting can be interpreted as few-shot retrieval: given a mini-batch, every example is iteratively used as a query, and the rest ones compose the gallery to search, i.e., the support set in few-shot setting. The rest examples are split into a positive set and a negative set. For every mini-batch, the learning objective of ranked list loss is to make the query closer to the positive set than to the negative set by a margin. Second, previous methods aim to pull positive pairs as close as possible in the embedding space. As a result, the intraclass data distribution tends to be extremely compressed. In contrast, we propose to learn a hypersphere for each class in order to preserve useful similarity structure inside it, which functions as regularisation. Extensive experiments demonstrate the superiority of our proposal by comparing with the state-of-the-art methods on the fine-grained image retrieval task. Our source code is available online: https://github.com/XinshaoAmosWang/Ranked-List-Loss-for-DML.

Weakly Supervised Deep Metric Learning for Community-Contributed Image Retrieval

Low rank metric learning for social image retrieval.

High-Rankness Regularized Semi-Supervised Deep Metric Learning for Remote Sensing Imagery

Distance Metric Learning from Uncertain Side Information with Application to Automated Photo Tagging

Sparse semantic metric learning for image retrieval

Deep Multimodal Distance Metric Learning Using Click Constraints for Image Ranking.

Robust Object Recognition Via Weakly Supervised Metric and Template Learning

Weakly Supervised Deep Matrix Factorization for Social Image Understanding.

Deep metric learning for image retrieval in smart city development

Multimodal-aware Weakly Supervised Metric Learning with Self-Weighting Triplet Loss

Deep Metric Learning Based on Scalable Neighborhood Components for Remote Sensing Scene Characterization

Community-Aware Photo Quality Evaluation by Deeply Encoding Human Perception

Ranked List Loss for Deep Metric Learning

Deep Relational Metric Learning.

Weakly-Supervised Deep Nonnegative Low-Rank Model for Social Image Tag Refinement and Assignment.

Deep Metric Learning Via Adaptive Learnable Assessment

Semantic Discriminative Metric Learning for Image Similarity Measurement

Hybrid-Attention Based Decoupled Metric Learning for Zero-Shot Image Retrieval

Online Progressive Deep Metric Learning

Multilevel Similarity-Aware Deep Metric Learning for Fine-Grained Image Retrieval

Metric Learning with Feature Decomposition for Image Categorization