Hardness-Aware Deep Metric Learning
This work addresses a bottleneck in metric learning for computer vision applications, but it is incremental as it builds on existing hard negative mining strategies.
The paper tackles the problem of insufficient informative samples in deep metric learning by adaptively generating synthetic embeddings to fully exploit all training data, achieving competitive performance on CUB-200-2011, Cars196, and Stanford Online Products datasets.
This paper presents a hardness-aware deep metric learning (HDML) framework. Most previous deep metric learning methods employ the hard negative mining strategy to alleviate the lack of informative samples for training. However, this mining strategy only utilizes a subset of training data, which may not be enough to characterize the global geometry of the embedding space comprehensively. To address this problem, we perform linear interpolation on embeddings to adaptively manipulate their hard levels and generate corresponding label-preserving synthetics for recycled training, so that information buried in all samples can be fully exploited and the metric is always challenged with proper difficulty. Our method achieves very competitive performance on the widely used CUB-200-2011, Cars196, and Stanford Online Products datasets.