Visual object tracking using adaptive correlation filters

Abstract

Although not commonly used, correlation filters can track complex objects through rotations, occlusions and other distractions at over 20 times the rate of current stateof-the-art techniques. The oldest and simplest correlation filters use simple templates and generally fail when applied to tracking. More modern approaches such as ASEF and UMACE perform better, but their training needs are poorly suited to tracking. Visual tracking requires robust filters to be trained from a single frame and dynamically adapted as the appearance of the target object changes. This paper presents a new type of correlation filter, a Minimum Output Sum of Squared Error (MOSSE) filter, which produces stable correlation filters when initialized using a single frame. A tracker based upon MOSSE filters is robust to variations in lighting, scale, pose, and nonrigid deformations while operating at 669 frames per second. Occlusion is detected based upon the peak-to-sidelobe ratio, which enables the tracker to pause and resume where it left off when the object reappears.

Tracking by Natural Language Specification

“…Many modern trackers rely on discriminative correlation filters [4,12,7]. While originally selected for the Fast Fourier Transform to compute one channel quickly, Danelljan et al [10] use multiple channels to augment the discrimination of the correlation filters.…”

Real-Time Object Tracking via Fusion of Global and Local Appearance Models

“…Therefore, it is reasonable to fuse them with equal weights. As similar to the previous well-known approaches [4], [5], [9], a typically used value for the learning rate is about τ = 0.05 ∼ 1.0 when we update the appearance incrementally in online object tracking. For this reason, we set r = 0.2; therefore, the maximum learning rate τ is bounded by 0.1.…”

“…The global representation of an object is robust against noise and background clutters due to the global geometric constraint and is computationally efficient, especially when we design a discriminative model. Among various methods that are based on the global representation, the correlation filter has been successfully applied to object tracking because of its computational efficiency in representing global statistical characteristic of object appearance against background [4], [5] despite of its limitation on scale variation and partial occlusion as shown in Fig. 1.…”

The Visual Object Tracking VOT2014 Challenge Results

“…The advances proposed by VOT have also influenced the development of related methodologies. For example, the recent [77] now acknowledges that their area under the curve is an average overlap measure and have also adopted a variant of resets from 7 http://www.votchallenge.net VOT. The recent [42] benchmark adapted the approach of analyzing performance on subsequences instead of entire sequences to study the effects of occlusion.…”