Deep Metric Learning via Lifted Structured Feature Embedding
This work addresses the challenge of metric learning for visual recognition, offering a method that enhances discriminative training in neural networks, though it is incremental as it builds on existing deep learning frameworks.
The paper tackles the problem of learning semantic feature embeddings for visual recognition by proposing a novel structured prediction objective that lifts pairwise distances to a matrix within training batches, achieving significant improvements over existing deep feature embedding methods on datasets like CUB-200-2011, CARS196, and a newly collected Online Products dataset.
Learning the distance metric between pairs of examples is of great importance for learning and visual recognition. With the remarkable success from the state of the art convolutional neural networks, recent works have shown promising results on discriminatively training the networks to learn semantic feature embeddings where similar examples are mapped close to each other and dissimilar examples are mapped farther apart. In this paper, we describe an algorithm for taking full advantage of the training batches in the neural network training by lifting the vector of pairwise distances within the batch to the matrix of pairwise distances. This step enables the algorithm to learn the state of the art feature embedding by optimizing a novel structured prediction objective on the lifted problem. Additionally, we collected Online Products dataset: 120k images of 23k classes of online products for metric learning. Our experiments on the CUB-200-2011, CARS196, and Online Products datasets demonstrate significant improvement over existing deep feature embedding methods on all experimented embedding sizes with the GoogLeNet network.