Generalization in Metric Learning: Should the Embedding Layer be the Embedding Layer?
This work addresses generalization issues in metric learning for fine-grained image retrieval, proposing a novel regularization practice that is incremental but impactful for the domain.
The paper tackles the problem of improving generalization in deep metric learning for fine-grained image retrieval by investigating the use of layers other than the embedding layer for feature extraction, resulting in state-of-the-art performance on three benchmarks: Cars-196, CUB-200-2011, and Stanford Online Product.
This work studies deep metric learning under small to medium scale data as we believe that better generalization could be a contributing factor to the improvement of previous fine-grained image retrieval methods; it should be considered when designing future techniques. In particular, we investigate using other layers in a deep metric learning system (besides the embedding layer) for feature extraction and analyze how well they perform on training data and generalize to testing data. From this study, we suggest a new regularization practice where one can add or choose a more optimal layer for feature extraction. State-of-the-art performance is demonstrated on 3 fine-grained image retrieval benchmarks: Cars-196, CUB-200-2011, and Stanford Online Product.