Label-Embedding for Image Classification

TL;DR

This paper introduces a label-embedding framework for image classification that improves zero-shot learning by embedding classes in attribute space and learning a compatibility function, outperforming standard methods.

Contribution

It proposes a novel label-embedding approach for image classification that leverages attribute vectors and can incorporate various information sources, enhancing zero-shot learning.

Findings

Outperforms standard attribute prediction in zero-shot scenarios

Leverages alternative information sources like hierarchies and descriptions

Applicable across different learning settings from zero-shot to large-data training

Abstract

Attributes act as intermediate representations that enable parameter sharing between classes, a must when training data is scarce. We propose to view attribute-based image classification as a label-embedding problem: each class is embedded in the space of attribute vectors. We introduce a function that measures the compatibility between an image and a label embedding. The parameters of this function are learned on a training set of labeled samples to ensure that, given an image, the correct classes rank higher than the incorrect ones. Results on the Animals With Attributes and Caltech-UCSD-Birds datasets show that the proposed framework outperforms the standard Direct Attribute Prediction baseline in a zero-shot learning scenario. Label embedding enjoys a built-in ability to leverage alternative sources of information instead of or in addition to attributes, such as e.g. class hierarchies or textual descriptions. Moreover, label embedding encompasses the whole range of learning settings from zero-shot learning to regular learning with a large number of labeled examples.