Scalable implementation of particle filter-based visual object tracking on network-on-chip (NoC)
- PDF / 3,170,661 Bytes
- 18 Pages / 595.276 x 790.866 pts Page_size
- 51 Downloads / 197 Views
ORIGINAL RESEARCH PAPER
Scalable implementation of particle filter‑based visual object tracking on network‑on‑chip (NoC) Pinalkumar Engineer1 · Rajbabu Velmurugan1 · Sachin Patkar1 Received: 24 February 2018 / Accepted: 3 December 2018 © Springer-Verlag GmbH Germany, part of Springer Nature 2019
Abstract Particle filter algorithms have been successfully used in various visual object tracking applications. They handle non-linear model and non-Gaussian noise, but are computationally demanding. In this paper, we propose a scalable implementation of particle filter algorithm for visual object tracking, using scalable interconnect such as network-on-chip on an FPGA platform. Here, several processing elements execute parallelly to handle large number of particles. We propose two designs and implementations, with one optimized for speed and other optimized for area. These implementations can easily support different image sizes, object sizes, and number of particles, without modifying the complete architecture. Multi-target tracking is also demonstrated for four objects. We validated the particle filter-based visual tracking with video feed from a Petalinux-based system. With image size of 320 × 240 , frame rates of 348 fps and 310 fps were achieved for single-object tracking of size 17 × 17 and 33 × 33 pixels, respectively, with a reasonable low-power consumption of 1.7 mW/fps on Zynq XC7Z020 (Zedboard) with an operating frequency of 69 MHz. This makes our implementation a good candidate for lowpower, visual object tracking using FPGA, especially in low-power, smart camera applications. Keywords Particle filter · Object tracking · FPGA · Network-on-chip (NoC) · Real-time image processing · Scalability · Low power · Smart camera
1 Introduction Object tracking is an important aspect of several computer vision applications such as robotic control or visual surveillance. It estimates object positions across frames in a video. A survey of popular object tracking algorithms can be found in [39, 41]. One successful approach is to pose this problem as a state estimation problem and use a Bayesian framework such as the particle filter (PF) algorithm [4, 22, 25, 38, 43]. Particle filter is an efficient algorithm to handle nonGaussian noise and non-linearity in state transition or measurement model. Particle filter algorithm represents the distributions in Bayesian estimation with a set of samples * Pinalkumar Engineer [email protected] Rajbabu Velmurugan [email protected] Sachin Patkar [email protected] 1
Electrical Engineering Department, Indian Institute of Technology Bombay, Mumbai, India
called particles and their corresponding weights. PF-based visual tracking was first proposed in [16]. A color-based particle filter tracker was proposed in [28] and was shown to outperform the mean-shift tracker in terms of reliability, at the price of increased computational time. One main concern for real-time particle filter (PF)-based visual tracking is to obtain results in a reasonable computational time with desired accu
Data Loading...
