TRUST: A Novel Framework for Vehicle Trajectory Recovery from Urban-Scale Videos.

We study a new type of path inference query against urban-scale video databases. Given a vehicle image query, our goal is to recover its historical trajectory from the footprints captured by surveillance cameras deployed across the road network. The problem is challenging because visual matching inherently suffers from object occlusion, low camera resolution, varying illumination conditions, and viewing angles. Furthermore, with limited computation resources, only a fraction of video frames can be ingested and indexed, causing severe data sparsity issues for visual matching. To support efficient and accurate trajectory recovery, we develop a select-and-refine framework in a heterogeneous hardware environment with both CPUs and GPUs. We construct a proximity graph from the top-k visually similar frames and propose holistic scoring functions based on visual and spatial-temporal coherence. To avoid enumerating all the paths, we also propose a coarse-grained scoring function with monotonic property to reduce search space. Finally, the derived path is refined by examining raw video frames to fill the missing cameras. For performance evaluation, we construct two largest-scale video databases generated from cameras deployed upon real road networks. Experimental results validate the efficiency and accuracy of our proposed trajectory recovery framework.

COVID-19 contact tracking by group activity trajectory recovery over camera networks.

Contact tracking plays an important role in the epidemiological investigation of COVID-19, which can effectively reduce the spread of the epidemic. As an excellent alternative method for contact tracking, mobile phone location-based methods are widely used for locating and tracking contacts. However, current inaccurate positioning algorithms that are widely used in contact tracking lead to the inaccurate follow-up of contacts. Aiming to achieve accurate contact tracking for the COVID-19 contact group, we extend the analysis of the GPS data to combine GPS data with video surveillance data and address a novel task named group activity trajectory recovery. Meanwhile, a new dataset called GATR-GPS is constructed to simulate a realistic scenario of COVID-19 contact tracking, and a coordinated optimization algorithm with a spatio-temporal constraint table is further proposed to realize efficient trajectory recovery of pedestrian trajectories. Extensive experiments on the novel collected dataset and commonly used two existing person re-identification datasets are performed, and the results evidently demonstrate that our method achieves competitive results compared to the state-of-the-art methods.

Wearable Wrist Movement Monitoring Using Dual Surface-Treated Plastic Optical Fibers.

Regarding high-sensitivity human wrist joint motion monitoring in exercise rehabilitation; we develop a pair of novel wearable and sensitivity-enhanced plastic optical fiber (POF) strain sensors consisting of an etched grating fiber and a side-polished fiber stitched into a polyamide wrist brace. The two flexible and surface-treated fibers are; respectively; featured with an etched periodic gratings with a pitch of 6 mm and a depth of 0.5 mm and a D-shaped side-polished zone of ~300 µm depth and ~30 mm length; which, correspondingly, show the sensitivities of around 0.0176/° and 0.0167/° in a normalized bending angle by far larger than a conventional commercial POF, because it achieves a more sensitive strain-induced evanescent field interaction with the side-machined fibers. Moreover, in terms of the sensor response to bending deformation in the range of -40°~+40°, the former exhibits a better sensitivity in lower angle change, while the latter is superior as the bending angle increases; thereby arranging the two modified POFs separately at the side and back of the human wrist, in order to decouple the wrist joint behaviors induced by typical flexion-extension or abduction-adduction movements. Then, the circular and pentagonal wrist motion trajectory patterns are investigated, to demonstrate the maximum average single-axis motion error of 2.94° via the transformation of spatial angle to plane coordinate for the fabricated couple of POF sensors, which is lower than a recognized standard of 5°, thus suggesting the great potential in wearable exercise rehabilitation of human joints in the field of medical treatment and healing.