Interestingness Prediction by Robust to Rank

Yanwei Fu, Timothy M. Hospedales, Tao Xiang, Yuan Yao and Shaogang Gong

School of EECS, Queen Mary University of London, UK {y.fu,t.hospedales, t.xiang, z.fu, s.gong}@qmul.ac.uk

Abstract

The problem of predicting image or video interestingness from their low-level feature representations has received increasing interest. As a highly subjective visual attribute, annotating the interestingness value of training data for learning a prediction model is challenging. To make the annotation less subjective and more reliable, recent studies employ crowdsourcing tools to collect pairwise comparisons – relying on majority voting to prune the annotation outliers/errors. In this paper, we propose a more principled way to identify annotation outliers by formulating the interestingness prediction task as a unified robust learning to rank problem, tackling both the outlier detection and interestingness prediction tasks jointly. Extensive experiments on both image and video interestingness benchmark datasets demonstrate that our new approach significantly outperforms state-of-the-art alternatives.

Download

Source codes: https://github.com/yanweifu/Robust_ranking

The website to collect the pairwise comparison of ages:  http://yanweifu.github.io/survey4/index.html

The website to collect the pairwise comparison of video relative attributes: http://yanweifu.github.io/usaasurvey/index.html  (We didnot use this dataset in our experiments&papers.)

Paper:

[1] Yanwei Fu, Timothy M. Hospedales, Tao Xiang, Yuan Yao and Shaogang Gong,:  Interestingness Prediction by Robust Learning to Rank,  (ECCV 2014).  Paper

bib: @INPROCEEDINGS{ranking2014ECCV,
author = { Yanwei Fu and Timothy M. Hospedales and Tao Xiang and Yuan Yao and Shaogang Gong},
title = {Interestingness Prediction by Robust Learning to Rank},
booktitle = {ECCV},
year = {2014}
}

[2] Yanwei Fu, Timothy M. Hospedales, Tao Xiang, Jiechao Xiong, Shaogang Gong, Yizhou Wang, and Yuan Yao:   Robust Subjective Visual Property Prediction from Crowdsourced Pairwise Labels,

IEEE TPAMI 2016, Vol 38, Issue 3. 

bib: @INPROCEEDINGS{ranking2014PAMI,
author = {Yanwei Fu and Timothy M. Hospedales and Tao Xiang and Jiechao Xiong and Yuan Yao and Shaogang Gong and Yizhou Wang},
title = {Robust Estimation of Subjective Visual Properties},
booktitle = { IEEE Transactions on Pattern Analysis and Machine Intelligence},
year = {2016}
}

[3]  Yanwei Fu, De-An Huang, Leonid Sigal. Robust Classification by Pre-conditioned LASSO and Transductive Diffusion Component Analysis  Arxiv paper:

bib: @INPROCEEDINGS{preLassoFu,
author = {Yanwei Fu and De-an Huang and Leonid Sigal},
title = {Robust Classification by Pre-conditioned LASSO and Transductive Diffusion Component Analysis},
booktitle = { arxiv: 1511.06340},
year = {2016}
}