Bayesian Joint Modelling for Object Localisation in Weakly Labelled Images

Z. Shi, T. M. Hospedales, T. Xiang

Research output: Contribution to journalArticlepeer-review

Abstract

We address the problem of localisation of objects as bounding boxes in images and videos with weak labels. This weakly supervised object localisation problem has been tackled in the past using discriminative models where each object class is localised independently from other classes. In this paper, a novel framework based on Bayesian joint topic modelling is proposed, which differs significantly from the existing ones in that: (1) All foreground object classes are modelled jointly in a single generative model that encodes multiple object co-existence so that “explaining away” inference can resolve ambiguity and lead to better learning and localisation. (2) Image backgrounds are shared across classes to better learn varying surroundings and “push out” objects of interest. (3) Our model can be learned with a mixture of weakly labelled and unlabelled data, allowing the large volume of unlabelled images on the Internet to be exploited for learning. Moreover, the Bayesian formulation enables the exploitation of various types of prior knowledge to compensate for the limited supervision offered by weakly labelled data, as well as Bayesian domain adaptation for transfer learning. Extensive experiments on the PASCAL VOC, ImageNet and YouTube-Object videos datasets demonstrate the effectiveness of our Bayesian joint model for weakly supervised object localisation.
Original languageEnglish
Pages (from-to)1959-1972
Number of pages14
JournalIEEE Transactions on Pattern Analysis and Machine Intelligence
Volume37
Issue number10
DOIs
Publication statusPublished - 1 Oct 2015

Fingerprint

Dive into the research topics of 'Bayesian Joint Modelling for Object Localisation in Weakly Labelled Images'. Together they form a unique fingerprint.

Cite this