Abstract:Discriminant correlation filter-based tracking approaches, which adopt a circular shift operator on the tracking target object (the only accurate positive sample) to obtain training data and rely on the potential sample periodic extension hypothesis that enables model training and detection, can be efficiently accomplished through FFT. However, real background information is not modeled during the total learning process. The background-aware correlation filter (BACF) tracking algorithm uses a binary matrix to acquire real positive and negative samples using a dense sampling method to model the targets appearance. However, the BACF algorithm does not consider temporal and spatial consistency information, and when a target undergoes an abrupt change, the learned correlation filter will drift to the background. To solve this problem, in this paper, we introduce temporal and spatial consistency constraints into the baseline BACF framework and propose a learning temporal-spatial consistency correlation filter (TSCF) tracking algorithm. This enables the correlation filter to learn to adapt to the appearance of mutation between successive frames. The temporal consistency constraint smooths the multi-channel correlation filter in the time series, and the spatial consistency constraint smooths the multi-channel correlation filter in spatial distribution, thus making the energy distribution more uniform of the correlation filter learned. In this paper, the TSCF model has closed solutions and the conjugate gradient descent method is used to approximate the optimal solution of a system of closed solutions. The optimization process can then be transformed into the Fourier domain using cyclic matrix properties to quickly obtain a solution, which effectively reduces the cost of calculating large matrices. In this paper, our TSCF algorithm increases distance precision by 5.5$%$ and raises the AUC by 4.3$%$ compared to the baseline BACF algorithm on the TB100 public database. The distance precision achieves 0.879 and the AUC reaches 0.663 on the TB100 database making use of only hand-crafted features. The TSCF algorithm proposed in this paper can be applied to challenging conditions such as short time occlusion, out-of-plane rotation, in-plane rotation, and so on, thus demonstrating its robustness and effectiveness.

Target Tracking Based on Improved STRCF Algorithm

High-speed Tracking with Multi-Templates Correlation Filters

Target tracker with masked discriminative correlation filter

Improved CACF Algorithm Based on Channel Reliability

Delayed rectification of discriminative correlation filters for visual tracking

Deep Correlation Filter based Real-Time Tracker

Learning Spatial-Temporal Regularized Correlation Filters for Visual Tracking

Real-Time Tracking Based on Keypoints and Discriminative Correlation Filters

Robust Visual Tracking Based On Adaptive Extraction And Enhancement Of Correlation Filter

Temporally-adjusted correlation filter-based tracking.

Collaborative Convolution Operators for Real-Time Coarse-to-Fine Tracking

Correlation filter tracking algorithm based on spatial-temporal regularization and context awareness

Robust Visual Tracking with Motion-Aware and Automatic Temporal Regularization

Robust visual tracking via spatio-temporal adaptive and channel selective correlation filters

An Improved Correlation Filter-Based Target Tracking Method

Discriminative Visual Tracking Via Spatially Smooth and Steep Correlation Filters

Faster Spatially Regularized Correlation Filters for Visual Tracking.

Learning Temporal-Spatial Consistency Correlation Filter for Visual Tracking

An Accelerated Correlation Filter Tracker

An Object Tracking Algorithm with Channel Reliability and Target Response Adaptation

Real-time long-term tracker with tracking–verification–detection–refinement