Learning and regularizing motion models for enhancing particle filter-based target tracking
PSIVT'11 Proceedings of the 5th Pacific Rim conference on Advances in Image and Video Technology - Volume Part II
Two-granularity tracking: mediating trajectory and detection graphs for tracking under occlusions
ECCV'12 Proceedings of the 12th European conference on Computer Vision - Volume Part V
SuperFloxels: a mid-level representation for video sequences
ECCV'12 Proceedings of the 12th international conference on Computer Vision - Volume Part III
Robust object tracking using constellation model with superpixel
ACCV'12 Proceedings of the 11th Asian conference on Computer Vision - Volume Part III
Human action recognition with salient trajectories
Signal Processing
Editor's Choice Article: Motion-based segmentation of objects using overlapping temporal windows
Image and Vision Computing
Shifted subspaces tracking on sparse outlier for motion segmentation
IJCAI'13 Proceedings of the Twenty-Third international joint conference on Artificial Intelligence
| Hi-index | 0.00 |
We propose a detection-free system for segmenting multiple interacting and deforming people in a video. People detectors often fail under close agent interaction, limiting the performance of detection based tracking methods. Motion information often fails to separate similarly moving agents or to group distinctly moving articulated body parts. We formulate video segmentation as graph partitioning in the trajectory domain. We classify trajectories as foreground or background based on trajectory saliencies, and use foreground trajectories as graph nodes. We incorporate object connectedness constraints into our trajectory weight matrix based on topology of foreground: we set repulsive weights between trajectories that belong to different connected components in any frame of their time intersection. Attractive weights are set between similarly moving trajectories. Information from foreground topology complements motion information and our spatiotemporal segments can be interpreted as connected moving entities rather than just trajectory groups of similar motion. All our cues are computed on trajectories and naturally encode large temporal context, which is crucial for resolving local in time ambiguities. We present results of our approach on challenging datasets outperforming by far the state of the art.