Scalable Nonlinear Embeddings for Semantic Category-based Image Retrieval

TL;DR

This paper introduces a scalable nonlinear embedding method for semantic image retrieval that efficiently learns from large datasets using a kernelized neural network approach, outperforming baselines on multiple datasets.

Contribution

The paper presents a novel scalable nonlinear embedding algorithm based on kernelization and neural networks, capable of handling large-scale training data efficiently.

Findings

Effective on seven challenging datasets.

Handles up to half a million training pairs.

Outperforms relevant baseline methods.

Abstract

We propose a novel algorithm for the task of supervised discriminative distance learning by nonlinearly embedding vectors into a low dimensional Euclidean space. We work in the challenging setting where supervision is with constraints on similar and dissimilar pairs while training. The proposed method is derived by an approximate kernelization of a linear Mahalanobis-like distance metric learning algorithm and can also be seen as a kernel neural network. The number of model parameters and test time evaluation complexity of the proposed method are O(dD) where D is the dimensionality of the input features and d is the dimension of the projection space - this is in contrast to the usual kernelization methods as, unlike them, the complexity does not scale linearly with the number of training examples. We propose a stochastic gradient based learning algorithm which makes the method scalable (w.r.t. the number of training examples), while being nonlinear. We train the method with up to half a million training pairs of 4096 dimensional CNN features. We give empirical comparisons with relevant baselines on seven challenging datasets for the task of low dimensional semantic category based image retrieval.