VrR-VG: Refocusing Visually-Relevant Relationships

Yuanzhi Liang, Yalong Bai, Wei Zhang, Xueming Qian, Li Zhu, Tao Mei; The IEEE International Conference on Computer Vision (ICCV), 2019, pp. 10403-10412

Abstract


Relationships encode the interactions among individual instances and play a critical role in deep visual scene understanding. Suffering from the high predictability with non-visual information, relationship models tend to fit the statistical bias rather than "learning" to infer the relationships from images. To encourage further development in visual relationships, we propose a novel method to mine more valuable relationships by automatically pruning visually-irrelevant relationships. We construct a new scene graph dataset named Visually-Relevant Relationships Dataset (VrR-VG) based on Visual Genome. Compared with existing datasets, the performance gap between learnable and statistical method is more significant in VrR-VG, and frequency-based analysis does not work anymore. Moreover, we propose to learn a relationship-aware representation by jointly considering instances, attributes and relationships. By applying the representation-aware feature learned on VrR-VG, the performances of image captioning and visual question answering are systematically improved, which demonstrates the effectiveness of both our dataset and features embedding schema. Both our VrR-VG dataset and representation-aware features will be made publicly available soon.

Related Material


[pdf]
[bibtex]
@InProceedings{Liang_2019_ICCV,
author = {Liang, Yuanzhi and Bai, Yalong and Zhang, Wei and Qian, Xueming and Zhu, Li and Mei, Tao},
title = {VrR-VG: Refocusing Visually-Relevant Relationships},
booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
month = {October},
year = {2019}
}