Combining Appearance and Range Based Information for Multi-class Generic Object Recognition

Authors:
Doaa Hegazy;Joachim Denzler
Affiliations:
Institute of Computer Science, Friedrich-Schiller-University in Jena, Jena, Germany 07743;Institute of Computer Science, Friedrich-Schiller-University in Jena, Jena, Germany 07743
Venue:
CIARP '09 Proceedings of the 14th Iberoamerican Conference on Pattern Recognition: Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications
Year:
2009

Citing 7
Cited 0

An Affine Invariant Interest Point Detector

ECCV '02 Proceedings of the 7th European Conference on Computer Vision-Part I
A New Paradigm for Recognizing 3-D Object Shapes from Range Data

ICCV '03 Proceedings of the Ninth IEEE International Conference on Computer Vision - Volume 2
Distinctive Image Features from Scale-Invariant Keypoints

International Journal of Computer Vision
Learning Generative Visual Models from Few Training Examples: An Incremental Bayesian Approach Tested on 101 Object Categories

CVPRW '04 Proceedings of the 2004 Conference on Computer Vision and Pattern Recognition Workshop (CVPRW'04) Volume 12 - Volume 12
Sharing Visual Features for Multiclass and Multiview Object Detection

IEEE Transactions on Pattern Analysis and Machine Intelligence
Object localization with boosting and weak supervision for generic object recognition

SCIA'05 Proceedings of the 14th Scandinavian conference on Image Analysis
Coloring local feature extraction

ECCV'06 Proceedings of the 9th European conference on Computer Vision - Volume Part II

Quantified Score

Hi-index	0.00

Visualization

Abstract

The use of range images for generic object recognition is not addressed frequently by the computer vision community. This paper presents two main contributions. First, a new object category dataset of 2D and range images of different object classes is presented. Second, a new generic object recognition model from range and 2D images is proposed. The model is able to use either appearance (2D) or range based information or a combination of both of them for multi-class object learning and recognition. The recognition performance of the proposed recognition model is investigated experimentally using the new database and promising results are obtained. Moreover, the best performance gain by combining both appearance and range based information is 35% for single classes while the average gain over classes is 12%.