The Group Loss++: A deeper look into group loss for deep metric learning

TL;DR

The paper introduces Group Loss++, a novel differentiable loss function for deep metric learning that improves clustering and retrieval by enforcing label consistency across all samples in a group, leading to state-of-the-art results.

Contribution

It proposes Group Loss++, a new loss function based on label propagation that considers all samples in a group, enhancing deep metric learning performance.

Findings

Achieves state-of-the-art clustering and retrieval results on four datasets.

Provides competitive results on person re-identification datasets.

Demonstrates the effectiveness of group-based loss over pair/triplet-based methods.

Abstract

Deep metric learning has yielded impressive results in tasks such as clustering and image retrieval by leveraging neural networks to obtain highly discriminative feature embeddings, which can be used to group samples into different classes. Much research has been devoted to the design of smart loss functions or data mining strategies for training such networks. Most methods consider only pairs or triplets of samples within a mini-batch to compute the loss function, which is commonly based on the distance between embeddings. We propose Group Loss, a loss function based on a differentiable label-propagation method that enforces embedding similarity across all samples of a group while promoting, at the same time, low-density regions amongst data points belonging to different groups. Guided by the smoothness assumption that "similar objects should belong to the same group", the proposed loss trains the neural network for a classification task, enforcing a consistent labelling amongst samples within a class. We design a set of inference strategies tailored towards our algorithm, named Group Loss++ that further improve the results of our model. We show state-of-the-art results on clustering and image retrieval on four retrieval datasets, and present competitive results on two person re-identification datasets, providing a unified framework for retrieval and re-identification.