Video cameras are pervasively deployed in city scale for public good or community safety (i.e. traffic monitoring or suspected person tracking). However, analyzing large scale video feeds in real time is data intensive and poses severe challenges to today's network and computation systems. We present CrossRoI, a resource-efficient system that enables real time video analytics at scale via harnessing the videos content associations and redundancy across a fleet of cameras. CrossRoI exploits the intrinsic physical correlations of cross-camera viewing fields to drastically reduce the communication and computation costs. CrossRoI removes the repentant appearances of same objects in multiple cameras without harming comprehensive coverage of the scene. CrossRoI operates in two phases - an offline phase to establish cross-camera correlations, and an efficient online phase for real time video inference. Experiments on real-world video feeds show that CrossRoI achieves 42% ~ 65% reduction for network overhead and 25% ~ 34% reduction for response delay in real time video analytics applications with more than 99% query accuracy, when compared to baseline methods. If integrated with SotA frame filtering systems, the performance gains of CrossRoI reaches 50% ~ 80% (network overhead) and 33% ~ 61% (end-to-end delay).

Nowadays, smartphones have become indispensable in people’s daily work and life. Since various sensors and communication chips have been integrated into smartphones, it has become feasible to provide indoor positioning using phones. This paper proposes such a solution based on a smartphone, combining Bluetooth low energy (BLE) and pedestrian dead reckoning (PDR) in the particle filter framework to realize real-time and reliable indoor positioning. First, the smartphone’s built-in accelerometer, magnetometer, and gyroscope are used to provide data measurements and formulate a feasible method for PDR. Second, a range-free weighted centroid algorithm is proposed to realize BLE-based localization with low computation complexity. However, a single positioning technology has limitations, e.g., the cumulative error of PDR and the received signal strength fluctuation of BLE. Finally, to exploit the complementary strengths of each technology, a fusion framework utilizing a particle filter is proposed to combine PDR and BLE-based methods and provides more stable and accurate positioning results. Experiments are conducted on a floor in a campus building. Experimental results show that our proposed fused positioning method offers more accurate and stable performance in the long run compared with single PDR or BLE-based positioning. The achieved average positioning error is 1.34 m, which is reduced by 24.16% compared with PDR positioning and 10.60% compared with BLE-based positioning. Moreover, about 95% of the positioning errors are smaller than 1.7 m. The proposed fused positioning method has a vast application prospect in indoor navigation, indoor user tracking, and interactive experience for indoor visitors.

Traditional direct SLAM methods formulate the camera pose and map estimation as minimization of the photometric error, which is tackled by the Gauss-Newton algorithm or the Levenberg-Marquardt algorithm in the optimization. However, the convexity of the photometric error only holds in a small region due to the characteristics of the convexity for grayscale images. Thus, the system may be stuck in sub-optimal local minima when tracking points have large displacement. Unlike grayscale images, the semantic probabilities omit the details inside the semantic objects while mainly reforming on the boundary of semantic objects, which has better convexity for large displacement. In this letter, we propose a novel semantic-direct visual odometry (SDVO), exploiting the direct alignment of semantic probabilities. By constructing the joint error function based on grayscale images and semantic probabilities, the joint error function achieves better convexity contrary to the photometric error. Consequently, the proposed system moves towards optima with steady steps in the optimization iterations. Experimental results on the challenging real-world dataset demonstrate a significant improvement over the baseline by integrating the direct alignment of semantic probabilities.

Human activities recognition (HAR) plays a vital role in fields like ambient assisted living and health monitoring, in which cross-subject recognition is one of the main challenges coming from the diversity of various users. Although recent studies have achieved satisfactory results in a non-cross-subject condition, the recognition performance has significant degradation under the cross-subject criterion. In this paper, we evaluate three traditional machine learning methods and five deep neural network architectures under the same metrics on three popular HAR datasets: mHealth, PAMAP2, and UCIDSADS. The experimental results show that traditional machine learning approaches are generally more robust to the new subject scenarios under strict leave-one-subject-out cross-validation. Extra analysis indicates that hand-crafted features are one major reason for the better performance of traditional machine learning on cross-subject HAR, while deep learning is more prone to learning subject-dependent features under an end-to-end training process. A novel training strategy for decision-tree-based methods is also proposed in this paper, resulting in an improvement on the random forest model which achieves competitive performance at an average F1-score (accuracy) of 94.49% (95.09%), 91.64% (92.21%), and 92.70% (93.29%) on the three datasets, compared with state-of-the-art solutions for cross-subject HAR.

In this letter, we propose a novel semantic direct monocular simultaneous localization and mapping (SLAM) system that fuses the semantic information obtained by an advanced deep neural network (DNN) into direct sparse odometry with loop closure(LDSO), with the purpose of improving the localization accuracy and building a dense semantic map of the urban environment. For localization, we apply a point reselection strategy based on coarse semantic plane (CSP) constraints to discard static points inconsistent with the nearby co-plane points of the same semantic class and dynamic points beyond the visible range. Moreover, a point group movement consistency (PGMC) check is utilized to decrease the impact of moving dynamic objects. For the dense semantic map, we model numerous small semantic planes from well-estimated points to measure the depth of each static pixel, rather than conduct stereo matching. Experimental results show that our method is more accurate than LDSO and comparable with ORB-SLAM in terms of localization. Moreover, it is capable of building a dense semantic map of the urban environment for better scene understanding.