An Improved Indirect Attribute Weighted Prediction Model for Zero-Shot Image Classification

Yuhu CHENG; Xue QIAO; Xuesong WANG

doi:10.1587/transinf.2015EDP7226

An Improved Indirect Attribute Weighted Prediction Model for Zero-Shot Image Classification

Yuhu CHENG, Xue QIAO, Xuesong WANG

Full Text Views

0

Cite this

Summary :

Zero-shot learning refers to the object classification problem where no training samples are available for testing classes. For zero-shot learning, attribute transfer plays an important role in recognizing testing classes. One popular method is the indirect attribute prediction (IAP) model, which assumes that all attributes are independent and equally important for learning the zero-shot image classifier. However, a more practical assumption is that different attributes contribute unequally to the classifier learning. We therefore propose assigning different weights for the attributes based on the relevance probabilities between the attributes and the classes. We incorporate such weighed attributes to IAP and propose a relevance probability-based indirect attribute weighted prediction (RP-IAWP) model. Experiments on four popular attributed-based learning datasets show that, when compared with IAP and RFUA, the proposed RP-IAWP yields more accurate attribute prediction and zero-shot image classification.

Publication: IEICE TRANSACTIONS on Information Vol.E99-D No.2 pp.435-442

Publication Date: 2016/02/01

Publicized: 2015/11/20

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2015EDP7226

Type of Manuscript: PAPER

Category: Pattern Recognition

Authors

Yuhu CHENG
  China University of Mining and Technology
Xue QIAO
  China University of Mining and Technology
Xuesong WANG
  China University of Mining and Technology

Keyword

indirect attribute prediction, relevance probability, attribute weight, zero-shot image classification

Cite this

Copy

Yuhu CHENG, Xue QIAO, Xuesong WANG, "An Improved Indirect Attribute Weighted Prediction Model for Zero-Shot Image Classification" in IEICE TRANSACTIONS on Information, vol. E99-D, no. 2, pp. 435-442, February 2016, doi: 10.1587/transinf.2015EDP7226.
Abstract: Zero-shot learning refers to the object classification problem where no training samples are available for testing classes. For zero-shot learning, attribute transfer plays an important role in recognizing testing classes. One popular method is the indirect attribute prediction (IAP) model, which assumes that all attributes are independent and equally important for learning the zero-shot image classifier. However, a more practical assumption is that different attributes contribute unequally to the classifier learning. We therefore propose assigning different weights for the attributes based on the relevance probabilities between the attributes and the classes. We incorporate such weighed attributes to IAP and propose a relevance probability-based indirect attribute weighted prediction (RP-IAWP) model. Experiments on four popular attributed-based learning datasets show that, when compared with IAP and RFUA, the proposed RP-IAWP yields more accurate attribute prediction and zero-shot image classification.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2015EDP7226/_p

Copy

@ARTICLE{e99-d_2_435,
author={Yuhu CHENG, Xue QIAO, Xuesong WANG, },
journal={IEICE TRANSACTIONS on Information},
title={An Improved Indirect Attribute Weighted Prediction Model for Zero-Shot Image Classification},
year={2016},
volume={E99-D},
number={2},
pages={435-442},
abstract={Zero-shot learning refers to the object classification problem where no training samples are available for testing classes. For zero-shot learning, attribute transfer plays an important role in recognizing testing classes. One popular method is the indirect attribute prediction (IAP) model, which assumes that all attributes are independent and equally important for learning the zero-shot image classifier. However, a more practical assumption is that different attributes contribute unequally to the classifier learning. We therefore propose assigning different weights for the attributes based on the relevance probabilities between the attributes and the classes. We incorporate such weighed attributes to IAP and propose a relevance probability-based indirect attribute weighted prediction (RP-IAWP) model. Experiments on four popular attributed-based learning datasets show that, when compared with IAP and RFUA, the proposed RP-IAWP yields more accurate attribute prediction and zero-shot image classification.},
keywords={},
doi={10.1587/transinf.2015EDP7226},
ISSN={1745-1361},
month={February},}

Copy

TY - JOUR
TI - An Improved Indirect Attribute Weighted Prediction Model for Zero-Shot Image Classification
T2 - IEICE TRANSACTIONS on Information
SP - 435
EP - 442
AU - Yuhu CHENG
AU - Xue QIAO
AU - Xuesong WANG
PY - 2016
DO - 10.1587/transinf.2015EDP7226
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E99-D
IS - 2
JA - IEICE TRANSACTIONS on Information
Y1 - February 2016
AB - Zero-shot learning refers to the object classification problem where no training samples are available for testing classes. For zero-shot learning, attribute transfer plays an important role in recognizing testing classes. One popular method is the indirect attribute prediction (IAP) model, which assumes that all attributes are independent and equally important for learning the zero-shot image classifier. However, a more practical assumption is that different attributes contribute unequally to the classifier learning. We therefore propose assigning different weights for the attributes based on the relevance probabilities between the attributes and the classes. We incorporate such weighed attributes to IAP and propose a relevance probability-based indirect attribute weighted prediction (RP-IAWP) model. Experiments on four popular attributed-based learning datasets show that, when compared with IAP and RFUA, the proposed RP-IAWP yields more accurate attribute prediction and zero-shot image classification.
ER -