A Scalable Platform for Distributed Object Tracking Across a Many-Camera Network

Abstract

Advances in deep neural networks (DNN) and computer vision (CV) algorithms have made it feasible to extract meaningful insights from large-scale deployments of urban cameras. Tracking an object of interest across the camera network in near real-time is a canonical problem. However, current tracking platforms have two key limitations: 1) They are monolithic, proprietary and lack the ability to rapidly incorporate sophisticated tracking models, and 2) They are less responsive to dynamism across wide-area computing resources that include edge, fog, and cloud abstractions. We address these gaps using Anveshak, a runtime platform for composing and coordinating distributed tracking applications. It provides a domain-specific dataflow programming model to intuitively compose a tracking application, supporting contemporary CV advances like query fusion and re-identification, and enabling dynamic scoping of the camera network's search space to avoid wasted computation. We also offer tunable batching and data-dropping strategies for dataflow blocks deployed on distributed resources to respond to network and compute variability. These balance the tracking accuracy, its real-time performance, and the active camera-set size. We illustrate the concise expressiveness of the programming model for four tracking applications. Our detailed experiments for a network of 1000 camera-feeds on modest resources exhibit the tunable scalability, performance, and quality trade-offs enabled by our dynamic tracking, batching, and dropping strategies.

A GPU Accelerated Dual-Ascent Algorithm for the Multidimensional Assignment Problem in a Multitarget Tracking Application

“…[27] proposed an approach to solving multi-assignment problems using successive oneto-one assignments of decreasing size with modified costs. References [28]- [31] and [32] proposed parallel algorithms for target tracking problems for various domains and applications. Reference [11] provided approximate solutions to the data association problem using dual decomposition and message passing algorithm, and [33] proposed a block ICM technique for assignment problems that converges to a local minimum.…”

Internet-of-Things Edge Computing Systems for Streaming Video Analytics: Trails Behind and the Paths Ahead

“…Scheduling algorithms are critical for optimizing the use of this complex environment, prompting researchers to explore more efficient algorithmic solutions. Some of the research works on edge video analytics, wherein scheduling is not the primary focus, have used simple approaches such as round-robin [82] and or bin-packing heuristics such as worst-fit [74]. Sophisticated approaches formulate the problem as a constraint optimization, and propose heuristics to solve it.…”

“…Many projects have used an embedded queuing library such as ZeroMQ for the data plane [73,77,82]. While embedded libraries perform well, they lack features such as data persistence, encryption, and replication, which are available in comprehensive data streaming frameworks like Apache Flink [101] and Apache Storm [102].…”

“…The Anveshak [82] project similarly exploits cross-camera correlations in pedestrian tracking. In contrast with the Spatula project, they use a physics-based approach, where the information on the road network and the speed of person movement are used to constrain the cameras that need to be activated.…”

Distributed System Load Balancing Strategy for Converged Workstations

“…The advantage of a single evaluation index system is that it is easy to code and implement, but there are certain defects. For example, when there are data blocks on a node server that are allocated for hotspot access, but its own disk space usage is relatively low, the load situation obtained by using only the disk space usage will be very different from the actual situation [7][8]. Through the above analysis, a single evaluation index cannot truly reflect the real load of each node in the DS.…”