Dual-Glance Model for Deciphering Social Relationships

ICCV 2017  ·  Junnan Li, Yongkang Wong, Qi Zhao, Mohan S. Kankanhalli ·

Since the beginning of early civilizations, social relationships derived from each individual fundamentally form the basis of social structure in our daily life. In the computer vision literature, much progress has been made in scene understanding, such as object detection and scene parsing. Recent research focuses on the relationship between objects based on its functionality and geometrical relations. In this work, we aim to study the problem of social relationship recognition, in still images. We have proposed a dual-glance model for social relationship recognition, where the first glance fixates at the individual pair of interest and the second glance deploys attention mechanism to explore contextual cues. We have also collected a new large scale People in Social Context (PISC) dataset, which comprises of 22,670 images and 76,568 annotated samples from 9 types of social relationship. We provide benchmark results on the PISC dataset, and qualitatively demonstrate the efficacy of the proposed model.

PDF Abstract ICCV 2017 PDF ICCV 2017 Abstract

Datasets

Introduced in the Paper:

PISC

Used in the Paper:

MS COCO Visual Genome PIPA
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Visual Social Relationship Recognition PIPA Dual-Glance Accuracy 59.6 # 3
Compare
Visual Social Relationship Recognition PIPA Pair CNN Accuracy 58.0 # 4
Compare
Accuracy (domain) 65.9 # 3
Compare
Visual Social Relationship Recognition PISC Dual-Glance mAP 63.2 # 5
Compare
mAP (Coarse) 79.7 # 4
Compare

Methods
Add Remove

No methods listed for this paper. Add relevant methods here
Contact us on: hello@paperswithcode.com . Papers With Code is a free resource with all data licensed under CC-BY-SA.
Terms Data policy Cookies policy