Significance of Softmax-based Features in Comparison to Distance Metric Learning-based Features
arXiv:1712.10151 · doi:10.1109/TPAMI.2019.2911075
Abstract
The extraction of useful deep features is important for many computer vision tasks. Deep features extracted from classification networks have proved to perform well in those tasks. To obtain features of greater usefulness, end-to-end distance metric learning (DML) has been applied to train the feature extractor directly. However, in these DML studies, there were no equitable comparisons between features extracted from a DML-based network and those from a softmax-based network. In this paper, by presenting objective comparisons between these two approaches under the same network architecture, we show that the softmax-based features perform competitive, or even better, to the state-of-the-art DML features when the size of the dataset, that is, the number of training samples per class, is large. The results suggest that softmax-based features should be properly taken into account when evaluating the performance of deep features.
arXiv comment6 pages
References in corpus (3)
Cited by in corpus (12)
- Hierarchy-based Image Embeddings for Semantic Image Retrieval
- Personalized Classifier for Food Image Recognition
- Unsupervised Embedding Learning via Invariant and Spreading Instance Feature
- Deep Learning vs. Traditional Computer Vision
- Multi-Representational Learning for Offline Signature Verification using Multi-Loss Snapshot Ensemble of CNNs
- A review of Recent Techniques for Person Re-Identification
- Semantic Drift Compensation for Class-Incremental Learning
- Recognition of Multiple Food Items in a Single Photo for Use in a Buffet-Style Restaurant
- Detecting Human-Object Interactions with Action Co-occurrence Priors
- DNN or k-NN: That is the Generalize vs. Memorize Question
- Generalization in Metric Learning: Should the Embedding Layer be the Embedding Layer?
- ACP++: Action Co-occurrence Priors for Human-Object Interaction Detection