Abstract:Visual tracking is a challenging task in computer vision. Recently correlation filter based trackers have gained much attention due to their high efficiency and impressive performance. Several methods have been developed to utilize deep features extracted from Convolutional Neural Networks (CNNs) for correlation filter tracking. Despite their success, most of these approaches suffer from low tracking speed due to high computation burden of the deep models. In this paper, we propose a CNN based real-time deep tracker for robust visual tracking. We first exploit hierarchical deep features as target representation, which can better distinguish the target from the backgrounds in the presence of complex situations. Before applied to learn correlation filters, the channel number of hierarchical features is reduced by the principal component analysis (PCA). This method can decrease both feature redundancy and computation. Then we construct a reliable map with features from the last convolutional layer as they encode the most semantic information. The reliable map is used to constraint the searching area where the target most likely exist. To further handle long-term tracking, we improve the model discrimination capability by introducing the original template into the current correlation filter model. As the target movement is usually small between two adjacent frames, we reuse the deep features from location step to update current model. The features are reused by making a circular shift with position displacement of the target, which increases the tracking speed significantly. Extensive experimental results on large benchmarks show that our proposed tracker can perform at real-time and achieves the state-of-the-art performance.

Learning Collaborative Sparse Correlation Filter for Real-Time Multispectral Object Tracking.

Collaborative Correlation Filters for Real-Time Tracking with Spatial Constraint.

Deep Correlation Filter based Real-Time Tracker

Visual Tracking with Sparse Correlation Filters

Robust Visual Tracking Via Multitask Sparse Correlation Filters Learning

Spatial-Spectral Weighted and Regularized Tensor Sparse Correlation Filter for Object Tracking in Hyperspectral Videos

Learning Adaptive Discriminative Correlation Filters via Temporal Consistency Preserving Spatial Feature Selection for Robust Visual Tracking

Learning spatial-channel regularization jointly with correlation filter for visual tracking

Learning Multi-Feature Based Spatially Regularized And Scale Adaptive Correlation Filters For Visual Tracking

Learning Low-Rank and Sparse Discriminative Correlation Filters for Coarse-to-Fine Visual Object Tracking

SCSTCF: Spatial-Channel Selection and Temporal Regularized Correlation Filters for visual tracking

Sparse Learning-Based Correlation Filter for Robust Tracking

Adaptive Correlation Filters with Long-Term and Short-Term Memory for Object Tracking

Sparse selective kernelized correlation filter model for visual object tracking

Aggregating Correlation Filter with Multiple Kernels Learning for Robust Visual Object Tracking.

Visual Object Tracking Via Collaborative Correlation Filters

Learning correlation filter with fused feature and reliable response for real-time tracking

Learning background-aware and spatial-temporal regularized correlation filters for visual tracking

Visual Tracking with Long-Short Term Based Correlation Filter

Exploiting the Anisotropy of Correlation Filter Learning for Visual Tracking

Long-Term Visual Tracking Based on Correlation Filters