Detecting Objects using Unsupervised Parts-based Attributes - Robotics Institute Carnegie Mellon University

Detecting Objects using Unsupervised Parts-based Attributes

Santosh Kumar Divvala, Charles Zitnick, Ashish Kapoor, and Simon Baker
Tech. Report, CMU-RI-TR-11-10, Robotics Institute, Carnegie Mellon University, August, 2010

Abstract

This paper presents a new approach to parts-based object detection. Objects are described using a spatial model based on its constituent parts. Unlike most existing methods, parts are discovered in an unsupervised manner from training images with only object bounding boxes provided. The association between parts is modeled using boosted decision trees that allows arbitrary object-part configurations to be maintained. Experimental results on the challenging VOC 2007 dataset validate our approach.

BibTeX

@techreport{Divvala-2010-10516,
author = {Santosh Kumar Divvala and Charles Zitnick and Ashish Kapoor and Simon Baker},
title = {Detecting Objects using Unsupervised Parts-based Attributes},
year = {2010},
month = {August},
institute = {Carnegie Mellon University},
address = {Pittsburgh, PA},
number = {CMU-RI-TR-11-10},
keywords = {Object Detection},
}