Unsupervised Modeling of Object Categories Using Link Analysis Techniques
People^Top
Gunhee Kim
Martial Hebert
Christos Faloutsos
Description^Top
This paper proposes a novel unsupervised modeling of object
categories by analyzing the statistics of the link structure (i.e., the
relationships between visual features). Given a set of
unlabelled images with a single number K, the number of object classes
to be classified, our approach solves the following two subtasks: (1)
Category discovery: clusters the input images into K groups according
to object categories. (2) Localization: detects the most probable
regions of the object in each classified image.
Our approach represents a set of visual information in the
form of a large-scale network (Fig.1.(a)). The
unsupervised classification and localization in visual modeling is
formulated as
the problem of finding hubs and communities in the network. The hubs
correspond to key class-specific visual features and the communities
map to
object categories. Following two link analysis techniques are applied
to infer those information: (1) PageRank
for ranked importance of visual features with respect to an image or a
category, (2) Blondel et al's vertex similarity algorithm for
structural similarity (Fig.1.(b)).
(a) A tiny part of the proposed network representation (b) Intution of
structural similarity
Figure 1. The
large-scale network representation for unsupervised modeling.
For category discovery, we observed that our approach achieved
competitive performance against most of previous work. For
localization, our approach fairly well discovered real class
representative features as hubs and suppressed the trivial
information as background (See Fig.2).
Figure 2. Examples of localization.
Publication^Top
Funding^Top
- Intelligent Robotics Development Program, a 21st Century Frontier R&D Programs by the Ministry of Commerce, Industry, and Energy of Korea.
Copyright notice^Top
The documents contained in these directories are included by the contributing authors as a means to ensure timely dissemination of scholarly and technical work on a non-commercial basis. Copyright and all rights therein are maintained by the authors or by other copyright holders, notwithstanding that they have offered their works here electronically. It is understood that all persons copying this information will adhere to the terms and constraints invoked by each author's copyright.