Label-Embedding for Attribute-Based Classification

Zeynep Akata 1, 2 Florent Perronnin 1 Zaid Harchaoui 2 Cordelia Schmid 2
2 LEAR - Learning and recognition in vision
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann, INPG - Institut National Polytechnique de Grenoble
Abstract : Attributes are an intermediate representation, which enables 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 which 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. The label embedding framework offers other advantages such as the ability to leverage alternative sources of information in addition to attributes (e.g. class hierarchies) or to transition smoothly from zero-shot learning to learning with large quantities of data.
Document type :
Conference papers
Complete list of metadatas
Contributor : Thoth Team <>
Submitted on : Friday, April 19, 2013 - 12:19:55 PM
Last modification on : Tuesday, February 12, 2019 - 10:30:05 AM
Long-term archiving on: Saturday, July 20, 2013 - 4:02:10 AM


Files produced by the author(s)


  • HAL Id : hal-00815747, version 1


Zeynep Akata, Florent Perronnin, Zaid Harchaoui, Cordelia Schmid. Label-Embedding for Attribute-Based Classification. IEEE Computer Vision and Pattern Recognition (CVPR), IEEE, Jun 2013, Portland, United States. ⟨hal-00815747v1⟩



Record views


Files downloads