L. S. Shapiro, H. Wang, and J. M. Brady. A matching and tracking strategy for independently moving objects. In D. Hogg and R. Boyle, editors, Proc. 3rd British Machine Vision Conf., Leeds, pages 306--315. Springer-Verlag, September 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....such as edges and corners over time. Essentially, what characterizes a feature tracking method is that image features are rst extracted in a bottom up processing step and then these features are used as the main primitives for the tracking and matching procedures. Concerning corner tracking, [Shapiro et al. 1992b] detect and track corners individually in an algorithm originally aimed at applications such as videoconferencing. Smith and Brady, 1995] track a large set of corners and use the results in a ow based segmentation algorithm. Zheng and Chellappa, 1995] have studied feature tracking when ....

....This measure is a normalized Gaussian weighted intensity cross correlation between two image patches. Here, we compute this measure over a square centered at the feature and with its size set from the detection scale. The measure is derived from the cross correlation of the image patches, see [Shapiro et al. 1992a] computed using a Gaussian weight function centered at the feature. The motivation for using a Gaussian weight function is that image structures near the feature center should be regarded as more signi cant than peripheral structures. Given two brightness functions I A and I B , and two image ....

L. S. Shapiro; H. Wang, and J. M. Brady. \A matching and tracking strategy for independently moving objects". In Proc. British Machine Vision Conference, pages 306-315. Springer Verlag, Berlin, 1992.

....and corners over time. Essentially, what characterizes a feature tracking method is that image features are rst extracted in a bottom up processing step and then these features are used as the main primitives for the tracking and matching procedures. Concerning corner tracking, Shapiro et al. [1] detect and track corners individually in an algorithm originally aimed at applications such as videoconferencing. Smith and Brady [2] track a large set of corners and use the results in a ow based segmentation algorithm. Zheng and Chellappa [3] have studied feature tracking when compensating for ....

L. S. Shapiro, H. Wang, and J. M. Brady, \A matching and tracking strategy for independently moving objects," in Proc. British Machine Vision Conference, pp. 306-315, Springer Verlag, Berlin, 1992.

....Following [2, 9, 17] we track a set of features through the sequence, and use their image motion to estimate the stabilizing warping. Other authors [8] use directly image intensities in a coarse to fine approach for single region tracking. We employ a modified version of the tracker described in [12], using a Kalman filter to predict feature s position. We adopt a fast outlier rejection rule (X84) in order to estimate the homography robustly. In this way, a moving object on a static background can be coped with. The tracking also takes advantage of the global warping computed at each frame, ....

....the ellipsoidal search region. If this matching fails (i.e. its normalized SSD is above a certain threshold) a search in a fixed neighborhood of the position in the previous frame is performed. If the matching still cannot be found, the feature goes into a particular state called ghost (after [12]) and it will be held as it was virtually still present for a short number of subsequent frames, after that either it reappears, or it is finally discarded. The duration of the ghost period must be chosen reasonably short (three frames in our case) if a feature disappears for a long time, it is ....

L. S. Shapiro, H. Wang, and J. M. Brady. A matching and tracking strategy for independently moving objects. In Proceedings of the British Machine Vision Conference, pages 306--315. BMVA Press, 1992.

....C = I 2 x I x I y I x I y I 2 y and I denotes the smoothing operation on I. Taking k equal to 0:04 and thresholding the result leads to corner detection. The implementation of the points tracker. The implementation has been strongly influenced by the corner tracker described in [7]. INRIA Applications of Non Metric Vision to Some Visual Guided Tasks 11 The correlation criterion used is: C(p 1 ; p 2 ) cos(i 1 Gamma i 1 ; i 2 Gamma i 2 ) i 1 Gamma i 1 ) i 2 Gamma i 2 ) ki 1 Gamma i 1 kki 2 Gamma i 2 k where i j is the vector of the local image intensity ....

H. Wang L. S. Shapiro and J. M. Brady. A matching and tracking strategy for independently-moving, non-rigid objects. In Proceedings of BMVC, 1992.

....Following [2, 9, 17] we track a set of features through the sequence, and use their image motion to estimate the stabilizing warping. Other authors [8] use directly image intensities in a coarse to fine approach for single region tracking. We employ a modified version of the tracker described in [12], using a Kalman filter to predict feature s position. We adopt a fast outlier rejection rule (X84) in order to estimate the homography robustly. In this way, a moving object on a static background can be coped with. The tracking also takes advantage of the global warping computed at each frame, ....

....the ellipsoidal search region. If this matching fails (i.e. its normalized SSD is above a certain threshold) a search in a fixed neighborhood of the position in the previous frame is performed. If the matching still cannot be found, the feature goes into a particular state called ghost (after [12]) and it will be held as it was virtually still present for a short number of subsequent frames, after that either it reappears, or it is finally discarded. The duration of the ghost period must be chosen reasonably short (three frames in our case) if a feature disappears for a long time, it ....

L. S. Shapiro, H. Wang, and J. M. Brady. A matching and tracking strategy for independently moving objects. In Proceedings of the British Machine Vision Conference, pages 306--315. BMVA Press, 1992.

.... of two dimensional features (such as corners) have the advantage that the full optical flow is known at every measurement position, because they do not suffer from the aperture problem effect (a discussion on this subject can be found in [24] Works on tracking of two dimensional features include [13, 1, 6, 18, 26]. Robust tracking means detecting automatically unreliable matches, or outliers, over an image sequence (see [14] for a survey of robust methods in computer vision) Recent examples of such robust algorithms include [23] which identifies tracking outliers while estimating the fundamental matrix, ....

L.S. Shapiro, H. Wang, and J.M. Brady. A matching and tracking strategy for independently moving objects. In Proceedings of the British Machine Vision Conference, pages 306--315. BMVA Press, 1992.

.... foveal ) window spatially and process at 25Hz with a latency of less than 70ms. For speed we used the detector of Wang and Brady [15, 1] rather than the Plessey detector used offline. Spatio temporal correspondence of corners was achieved using simple variation [12] of the algorithms proposed in [2, 13]. An important feature of the variant is that the image motion of the corners induced by the movement of the camera is calculated for each frame using odometric information and subtracted from that observed. Vision Controller Platform Motors Visual demand Visual feedback Encoder feedback Open ....

L. S. Shapiro, H. Wang, and J. M. Brady. A matching and tracking strategy for independently moving objects. In D. Hogg and R. Boyle, editors, Proc. 3rd British Machine Vision Conf., Leeds, pages 306--315. Springer-Verlag, September 1992.

.... In fact methods which extract explicit temporal structure (rather than independent motion estimates over time) are based on the matching and tracking of segmented features such as corners and employ motion models which, unlike affine motion models, contain temporal components in their motion models[5, 10, 11]. The method proposed in this paper addresses the problem of temporal independence. Motion results from one frame are used to guide the estimation of motion in the next. Specifically, previous optical flow fields are fed forward (actually warped forward) to act as initial estimates in an affine ....

L.S. Shapiro, H. Wang, and J.M. Brady. "A Matching and Tracking Strategy for Independently Moving Objects". In British Machine Vision Conference, pages 306--315, Leeds, UK, September 1992. Springer-Verlag.

....and T 1;2 are thresholds on edge and corner strengths. The second order properties are computed by linear interpolation between first order properties. Correspondence of corners between frames is achieved using a variation of the algorithms proposed by Shapiro, Wang and Brady (Brady et al. 1992; Shapiro et al. 1992). Corners are tracked from frame to frame using a constant image velocity Kalman filter (Bar Shalom and Fortmann 1988) state vector: x(k) x(k) y(k) x(k) y(k) state transition: x(k 1) Fx(k) v(k) where v(k) are Gaussian, zero mean, temporally uncorrelated noises with covariance ....

L. S. Shapiro, H. Wang, and J. M. Brady. A matching and tracking strategy for independently moving objects. In D. Hogg and R. Boyle, editors, Proc. 3rd British Machine Vision Conf., Leeds, pages 306--315. SpringerVerlag, September 1992.

....of them is the case of an object that is spinning much faster than the camera system can roll about its optical axis) It should be pointed out that we propose algorithms for an approximation of a very nonlinear problem. The major differences of our system from similar research efforts [3] 4] 5][19][20] 21] are the use of a single moving camera, the ability to compensate for inaccurate camera parameters and unknown depth (distance from the camera to the target) full 3D tracking ability, the small number of parameters that are estimated on line, and the integration of the characteristics of ....

L.S. Shapiro, H. Wang, and J.M. Brady, "A Matching and Tracking Strategy for Independently Moving Objects," in Proc. of the British Machine Vision Conference, pp. 306-315, 1992.

....of neighbors with respect to the current node. Even though all these constraints favor local rigidity, the particular image structure permits global deformation between image sequences. This technique has been applied to track fluid flow features in combustion image sequences [60] Shapiro et al. [56] presented a parallel strategy for tracking corner features on independently moving (and possibly non rigid) objects. Their system consists of two components: a matcher and a tracker. The matcher computes the correspondence based on local patch correlation, while the tracker supervises the matcher ....

L. S. Shapiro, H. Wang, and J. M. Brady. A matching and tracking strategy for independently moving objects. Proc. British Machine Vision Conference, pages 306--315, 1992.

....3.3 The PUMA Sequence 3 Figure (2) shows those trajectories that were tracked from frame 1 of the PUMA sequence in which the camera undergoes a rotational motion. Only trajectories of length greater than 6 are displayed and the square and circle symbols denote the start and 2 Shapiro et al. [16] call this the product moment coefficient. They point out that such a measure is invariant to linear changes in intensity and therefore compares the structure of the patches rather than their absolute intensities. 3 Puma and Toycar sequences are courtesy of the University of Massachusetts. ....

L. S. Shapiro, H. Wang, and J. M. Brady. A matching and tracking strategy for independently moving objects. In Proc. British Machine Vision Conference, pages 306--315, 1992.

No context found.

L. S. Shapiro, H. Wang, and J. M. Brady. A matching and tracking strategy for independently moving objects. In D. Hogg and R. Boyle, editors, Proc. 3rd British Machine Vision Conf., Leeds, pages 306--315. Springer-Verlag, September 1992.

No context found.

Larry S. Shapiro, Han Wang, and J. Michael Brady. A matching and tracking strategy for independently moving objects. In British Machine Vision Conference, pages 306--315, 1992.

No context found.

L. S. Shapiro, H. Wang, and J. M. Brady. A matching and tracking strategy for independently moving objects. In Proceedings of the British Machine Vision Conference, pages 308--315, 1992.

No context found.

L. S. Shapiro, H. Wang, and J. M. Brady. A matching and tracking strategy for independently moving objects. In D. Hogg and R. Boyle, editors, Proc. 3rd British Machine Vision Conf., Leeds, pages 306--315. Springer-Verlag, September 1992.

No context found.

Larry S. Shapiro, Han Wang, and J. Michael Brady. A matching and tracking strategy for independently moving objects. In British Machine Vision Conference, pages 306--315, 1992.

No context found.

Shapiro LS, Wang H, Brady JM. A matching and tracking strategy for independently moving objects. Proceedings of the British Machine Vision Conference, BMVA Press, 1992; 306-- 315

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