Label-Embedding for Image Classification
This addresses the problem of image classification with scarce training data for researchers in computer vision, offering a flexible framework that can leverage various information sources like attributes or hierarchies.
The paper tackles zero-shot image classification by framing attribute-based classification as a label-embedding problem, where classes are embedded in attribute vector spaces and a compatibility function ranks correct classes higher; results show it outperforms the Direct Attribute Prediction baseline on Animals With Attributes and Caltech-UCSD-Birds datasets.
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.