ixez / zkcf

zKCF is an extensible C++ implementation of KCF visual tracker.

zKCF

zKCF is an extensible C++ implementation of KCF(Kernelized Correlation Filters) visual tracker.

This project is mainly based on the code of KCFcpp and KCF Matlab[1][2]. In zKCF, the implementation of KCF's main body, feature extractors and correlation kernels are seperated, implementation is refined and reorganized for code readability and extensibility. Moreover, zKCF obtain a better performance and a faster speed with refined implementation and VGG feature extractor.

Evaluation and Comparison

The performance and speed of zKCF and its base KCFcpp are evaluated and compared on CVPR13[3] and OTB50/100[4] tracking benchmarks. zKCF_Vgg exploits VGG16's conv5_1 layer as the feature extractor.

Performance

Speed (FPS)

	zKCF_Vgg	zKCF_Hog	zKCF_HogLab	KCFcpp_Hog	KCFcpp_HogLab
CVPR13	12.23	90.48	52.70	102.89	76.54
OTB50	12.54	99.80	56.17	103.47	75.58
OTB100	12.49	102.63	57.52	107.44	78.05

Demo Usage

This section is deprecated due to the new dependencies of Caffe, Boost and Glog. Please read CMakeLists.txt and the relative files. More details will be supplemented soon.

Compilation

zKCF's dependencies include CMake(>3.0) and OpenCV(2/3).
Compilation follows an ordinary procedure of CMake project and is tested under Ubuntu 16.04.

mkdir build
cd build
cmake ..
make -j

Ubuntu 14.04 and older Linux OSs may have older CMake(2.x) and need to be updated. Windows should cooperate with Visual Studio and configure the dependencies. MacOS is expected to have a smooth compilation as under Ubuntu 16.04.

Run

The main function in Run.cpp is default for OTB[3][4] datasets and should be called as:

./zKCF seq_name /path/to/seq/ start_frame end_frame zero_padding ext bbox_x bbox_y bbox_width bbox_height preview

A Basketball sequence is prepared in assets/seqs for demo. The demo is called as:

./zKCF car4 /home/zeke/Documents/OTB/CVPR13/Car4/img/ 1 659 4 jpg 69 50 107 87 1

Development

Project Structure

• src/Run.cpp: main function.
• src(include)/KCF: Main class of zKCF.
• include/Def.h: Definitions of common constant variables.
• src(include)/Features: Feature extractors. All implement the interface IFeature.h.
  • RawFeature: To use raw pixels as features.
  • HogFeature: HoG feature extractor.
• src(include)/Kernels: Correlation kernels. All implement the interface IKernel.h.
  • GaussianKernel: Gaussian correlation kernel.
• src(include)/FkFactory: Factory class, generating and configuring different features and kernels.

How to extend zKCF

To add a new feature or kernel,

1. Put implementation codes in corresponding directories src(include)/Features and src(include)/Kernels.
2. Define a new FeatureType or KernelType in include/Def.h.
3. Customize parameters initialization in src/FkFactory.cpp.

Since features and kernels are based on interface IFeature.h and IKernel.h, implementation details are hidden in KCF class, which focusing on the pipeline of the tracker instead of features and kernels.
Nevertheless, parameters can be initialized for different features and kernels in ParamsInit method of KCF class.

TODOs

• Features
  • HogLabFeature
  • VGG Feature (conv5)
  • VGG Feature (conv1)
• Kernels
  • Linear kernel
  • Polynomial kernel

References

[1] Henriques, J. F., et al. "High-Speed Tracking with Kernelized Correlation Filters." IEEE Transactions on Pattern Analysis & Machine Intelligence 37.3(2015):583-596.
[2] Rui, Caseiro, P. Martins, and J. Batista. "Exploiting the circulant structure of tracking-by-detection with kernels." European Conference on Computer Vision Springer-Verlag, 2012:702-715.
[3] Wu, Yi, J. Lim, and M. Yang. "Online Object Tracking: A Benchmark Supplemental Material." 9.4(2013):2411-2418.
[4] Wu, Yi, Jongwoo Lim, and Ming-Hsuan Yang. "Object tracking benchmark." IEEE Transactions on Pattern Analysis and Machine Intelligence 37.9 (2015): 1834-1848.