| H.S. Parry, A.D. Marshall, and K.C. Markham. Integration of segmentation information and correlation technique for tracking objects in sequences of images. In Proc. of Videometrics IV (SPIE Proceedings vol. 2598), Philadelphia, PA, October 1995, pp. 208--219. |
....zooming for scale change. Changing scale poses increased demands on target identification and localization because commonly used approaches such as cross correlation are not scale invariant. Alternatives include feature based tracking (e.g. 17] or the use of adaptive correlation templates (e.g. [12]) 2.4 National Institute of Standards and Technology (Martin Herman, David Coombs, Sandor Szabo, Tsai Hong Hong) NIST was responsible for developing the vision processing platform, assisting in integrating University software onto the platform, and running the platform on vehicles at the NIST ....
H.S. Parry, A.D. Marshall, and K.C. Markham. Integration of segmentation information and correlation technique for tracking objects in sequences of images. In Proc. of Videometrics IV (SPIE Proceedings vol. 2598), Philadelphia, PA, October 1995, pp. 208--219.
....cope only with incremental changes in the appearance of the tracked object. A full search over the space of all template variations becomes unnecessary; instead, the template can be modified stepwise to keep pace with the appearance changes [Jackson, 1981; Tam et al. 1990; Montera et al. 1994; Parry et al. 1995] Adaptive correlation is therefore our choice for the task of tracking with varying focal length. There are two possibilities for updating the template. It can be warped resampled, or it can be replaced by the matched image patch. Warping the template has the advantage that it does not cause ....
....ways to reduce template drift have been proposed. In [Jackson, 1981] the template is only updated when the correlation score falls below a fixed threshold. In infrared imagery, thresholding the image provides target boundaries on which the new template can be recentered [Montera et al. 1994; Parry et al. 1995] Our strategy for updating the template is based on measuring the scaling between the template and its match in the current frame. Only when the scaling becomes large enough to impair the correlation, the template is updated. The next section introduces a differential method for estimating the ....
H. S. Parry, A. D. Marshall, and K. C. Markham. Integration of segmentation information and correlation technique for tracking objects in sequences of images. In Videometrics IV, SPIE Proceedings vol. 2598, pages 208--219, Philadelphia, PA, 25-- 26 October 1995.
....Changing scale poses increased demands on target identification and localization because commonly used approaches such as cross correlation are not scale invariant. Alternatives include feature based tracking (e.g. Reid and Murray, 1993 ] or the use of adaptive correlation templates (e.g. Parry et al. 1995 ] We are currently investigating the latter approach. Target distance m sec 1.30 1.40 1.50 1.60 1.70 1.80 0.00 5.00 10.00 15.00 20.00 Focal length mm sec 19.00 20.00 21.00 22.00 23.00 24.00 0.00 5.00 10.00 15.00 20.00 Figure 6: Focal length responding to changing target distance. 2.4 NIST ....
H.S. Parry, A.D. Marshall, and K.C. Markham. Integration of segmentation information and correlation technique for tracking objects in sequences of images. In Proc. of Videometrics IV (SPIE Proceedings vol. 2598), Philadelphia, PA, October 1995, pp. 208--219.
....Changing scale poses increased demands on target identification and localization because commonly used approaches such as cross correlation are not scale invariant. Alternatives include feature based tracking (e.g. Reid and Murray, 1993 ] or the use of adaptive correlation templates (e.g. Parry et al. 1995 ] We are currently investigating the latter approach. 2.4 NIST NIST is responsible for developing the vision processing platform, assisting in integrating University software onto the platform, and running the platform on vehicles at the NIST facility. In addition, NIST has collected video ....
H.S. Parry, A.D. Marshall, and K.C. Markham. Integration of segmentation information and correlation technique for tracking objects in sequences of images. In Proc. of Videometrics IV (SPIE Proceedings vol. 2598), Philadelphia, PA, October 1995, pp. 208--219.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC