TR2008-031

Learning on Lie Groups for Invariant Detection and Tracking


TR Image
The effect of model update. The target has severe appearance variations due to illumination and pose change throughout the sequence.
Abstract:

This paper presents a novel learning based tracking model combined with object detection. The existing technique proceed by linearizing the motion, which makes an implicit Euclidean space assumption. Most of the transformations used in computer vision have matrix Lie group structure. We learn the motion model on the Lie algebra and show that the formulation minimizes a first order approximation to the geodesic error. The learning model is extended to train a class specific tracking function, which is then integrated to an existing pose dependent object detector to build a pose invariant object detection algorithm. The proposed model can accurately detect objects in various poses, where the size of the search space is only a fraction compared to the existing object detection methods. The detection rate of the original detector is improved by more than 90% for large transformations.

 

  • Related News & Events

    •  NEWS    CVPR 2008: 8 publications by Jay Thornton, Shantanu D. Rane, Oncel Tuzel, Matthew Brand, Anthony Vetro and Amit K. Agrawal
      Date: June 28, 2008
      Where: IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
      MERL Contacts: Matthew Brand; Anthony Vetro
      Brief
      • The papers "Non-Refractive Modulators for Encoding and Capturing Scene Appearance and Depth" by Veeraraghavan, A., Agrawal, A., Raskar, R., Mohan, A. and Tumblin, J., "Feature Transformation of Biometric Templates for Secure Biometric Systems based on Error Correcting Codes" by Sutcu, Y., Rane, S., Yedidia, J.S., Draper, S.C. and Vetro, A., "Constant Time O(1) Bilateral Filtering" by Porikli, F., "Learning on Lie Groups for Invariant Detection and Tracking" by Tuzel, O., Porikli, F. and Meer, P., "Kernel Integral Images: A Framework for Fast non-Uniform Filtering" by Hussein, M., Porikli, F. and Davis, L., "A Conditional Random Field for Automatic Photo Editing" by Brand, M. and Pletscher, P., "Boosting Adaptive Linear Weak Classifiers for Online Learning and Tracking" by Parag, T., Porikli, F. and Elgammai, A. and "Sensing Increased Image Resolution Using Aperture Masks" by Mohan, A., Huang, X., Tumblin, J. and Raskar, R. were presented at the IEEE Conference on Computer Vision and Pattern Recognition (CVPR).
    •