Real-Time Embedded Object Tracking with Discriminative Correlation Filters Using Convolutional Features

Abstract

Object tracking is an essential element of visual perception systems. It is used in advanced video surveillance systems (AVSS), autonomous vehicles, robotics, and many more. For applications such as autonomous robots, the system must be implemented on some embedded platform with limited computing performance and power. Furthermore, sufficiently fast response is required from the tracking system in order to perform some real-time tasks. Discriminative Correlation Filter (DCF) based tracking algorithms are popular for such applications, as they offer state-of-the-art performance while not being too computationally complex. In this paper, an FPGA implementation of the DCF tracking algorithm using convolutional features is presented. The ZCU104 board is used as a platform, and the performance is evaluated on the VOT2015 dataset. In contrast to other implementations that use HOG (Histogram of Oriented Gradients) features, this implementation achieves better results for 64 × 64 filter size while being able to potentially operate at higher speeds (over 467 fps per scale).