Groupwise Tracking of Crowded Similar-Appearance Targets from Low-Continuity Image Sequences Hongkai Yu*, Youjie Zhou*, Jeff Simmons, C. Przybyla, Yuewei Lin, Xiaochuan Fan,
Yang Mi and Song Wang (* Co-first authors) CVPR 2016 Spotlight Oral
Automatic tracking of large-scale crowded targets are of particular importance in many applications, such as crowded people/vehicle tracking in video surveillance, fiber tracking in materials science, and cell tracking in biomedical imaging. This problem becomes very challenging when the targets show similar appearance and the interslice/inter-frame continuity is low due to sparse sampling, camera motion and target occlusion. The main challenge comes from the step of association which aims at matching the predictions and the observations of the multiple targets. In this paper we propose a new groupwise method to explore the target group information and employ the within-group correlations for association and tracking. In particular, the within-group association is modeled by a nonrigid 2D Thin-Plate transform and a sequence of group shrinking, group growing and group merging operations are then developed to refine the composition of each group. We apply the propose method to track large-scale fibers from the microscopy material images and compare its performance against several other multi-target tracking methods. We also apply the proposed method to track crowded people from videos with poor inter-frame continuity.
In the experiments, we apply the proposed method to track large-scale fibers from S200, an amorphous SiNC matrix reinforced by continuous Nicalon fibers. Three sets of data are collected, each of which is a 100-slice image sequence with dense inter-slice distance 1 μm. The image resolution of each slice is 1292-by-968. A sample slice is shown in Fig. 2(a), which contains hundreds of crowded fibers. On the collected data, we annotate the locations of fibers on each slice and link them across slices as the ground truth for performance evaluation. To test the tracking performance under sparsely sampled image sequences, we downsample the original image sequence. In particular, we skip C >= 0 slices before taking the next slice in the original sequence, until the end of original sequence is reached, to construct such sparsely sampled image sequences. For convenience, we name parameter C the sparsity: The larger the parameter C, the lower the interslice continuity of the constructed image sequence.[DataSet (image, detection, ground truth, sampled image list)] [TrackCheck Visualization Tool (Win, Linux, Mac)]