Sentence examples for motion discontinuity detection from inspiring English sources

As these contours are computed at high spatial resolution, they help to find the exact local position of the boundary for an object detected by MSTlModel center-surround neurons for motion discontinuity detection.

Here, the segmentation after the interaction between motion discontinuity detection and occlusion detection is shown.

The results show the successful occlusion and motion discontinuity detection of a real sequence with translational self motion and objects at different distances.

Our results show that the segmentation of moving objects can be considerably improved if occlusion and motion discontinuity detection mutually interact.

Furthermore, we show that the idea of temporal integration for these regions is again both for the occlusion and the motion discontinuity detection - a mechanism to get more robust results.

Other relevant insights are the application of different weights to each of the smoothness terms to add robustness against motion discontinuities, the detection of occluded regions (Ayvaci et al., 2010) and the application of a smoothing filter to the flow after each iteration of the optimization procedure (Sun et al., 2010; Thirion, 1998) to improve accuracy.

We propose a biologically inspired model for object segmentation that includes processing components for motion detection, and in contrast with previous approaches, makes use of both motion discontinuity and occlusion detection.

In the following subsection, we will a) show the results for motion discontinuity and occlusion detection, b) provide examples for object segmentation and estimation of the ordinal depth order, and c) demonstrate the effects of interactions between MSTlModel and TOModel.

Mutual interactions between motion discontinuities and occlusion/disocclusion detection based on temporal center-surround competition can be applied to stabilize boundary detection.

Hence, no occlusion region can be found that is adjacent to a detected and grouped motion discontinuity, this is a strong hint for a false detection.

In contrast, motion discontinuities are detected all along the outline of the tree.