文件列表:

KCF\choose_video.m (897, 2014-05-03)
KCF\download_videos.m (1749, 2014-05-03)
KCF\external.txt (1986, 2014-06-24)
KCF\fhog.m (3169, 2014-05-03)
KCF\gaussian_correlation.m (939, 2014-05-03)
KCF\gaussian_shaped_labels.m (1292, 2014-05-03)
KCF\get_features.m (1237, 2018-11-14)
KCF\get_subwindow.m (806, 2014-05-03)
KCF\gradientMex.mexa64 (23054, 2013-09-29)
KCF\gradientMex.mexw64 (30720, 2014-01-28)
KCF\linear_correlation.m (568, 2014-05-03)
KCF\load_video_info.m (3132, 2014-05-03)
KCF\polynomial_correlation.m (799, 2014-05-03)
KCF\precision_plot.m (1457, 2018-11-14)
KCF\run_KCF.m (1636, 2019-01-02)
KCF\run_tracker.m (7367, 2018-12-31)
KCF\show_video.m (2849, 2018-11-14)
KCF\tmpOPE.txt (1014, 2018-12-31)
KCF\tracker.m (5887, 2014-05-03)
KCF\videofig.m (7692, 2013-12-27)
KCF\results\results_1OPE_CVPR18 (0, 2018-12-30)
KCF\tmp\OPE (0, 2018-12-30)
KCF\results (0, 2018-12-30)
KCF\tmp (0, 2018-12-30)
KCF (0, 2018-12-30)

High-Speed Tracking with Kernelized Correlation Filters J. F. Henriques R. Caseiro P. Martins J. Batista TPAMI 2014 ________________ To be published. arXiv pre-print: http://arxiv.org/abs/1404.7584 Project webpage: http://www.isr.uc.pt/~henriques/circulant/ This MATLAB code implements a simple tracking pipeline based on the Kernelized Correlation Filter (KCF), and Dual Correlation Filter (DCF). It is free for research use. If you find it useful, please acknowledge the paper above with a reference. __________ Quickstart 1. Extract code somewhere. 2. The tracker is prepared to run on any of the 50 videos of the Visual Tracking Benchmark [3]. For that, it must know where they are/will be located. You can change the default location 'base_path' in 'download_videos.m' and 'run_tracker.m'. 3. If you don't have the videos already, run 'download_videos.m' (may take some time). 4. Execute 'run_tracker' without parameters to choose a video and test the KCF on it. Note: The tracker uses the 'fhog'/'gradientMex' functions from Piotr's Toolbox. Some pre-compiled MEX files are provided for convenience. If they do not work for your system, just get the toolbox from http://vision.ucsd.edu/~pdollar/toolbox/doc/index.html __________ The main interface function is 'run_tracker'. You can test several configurations (KCF, DCF, MOSSE) by calling it with different commands: run_tracker Without any parameters, will ask you to choose a video, track using the Gaussian KCF on HOG, and show the results in an interactive figure. Press 'Esc' to stop the tracker early. You can navigate the video using the scrollbar at the bottom. run_tracker VIDEO Allows you to select a VIDEO by its name. 'all' will run all videos and show average statistics. 'choose' will select one interactively. run_tracker VIDEO KERNEL Choose a KERNEL. 'gaussian'/'polynomial' to run KCF, 'linear' for DCF. run_tracker VIDEO KERNEL FEATURE Choose a FEATURE type, either 'hog' or 'gray' (raw pixels). run_tracker(VIDEO, KERNEL, FEATURE, SHOW_VISUALIZATION, SHOW_PLOTS) Decide whether to show the scrollable figure, and the precision plot. Useful combinations: >> run_tracker choose gaussian hog %Kernelized Correlation Filter (KCF) >> run_tracker choose linear hog %Dual Correlation Filter (DCF) >> run_tracker choose gaussian gray %Single-channel KCF (ECCV'12 paper) >> run_tracker choose linear gray %MOSSE filter (single channel) For the actual tracking code, check out the 'tracker' function. Though it's not required, the code will make use of the MATLAB Parallel Computing Toolbox automatically if available. __________ References [1] J. F. Henriques, R. Caseiro, P. Martins, J. Batista, "High-Speed Tracking with Kernelized Correlation Filters", TPAMI 2014 (to be published). [2] J. F. Henriques, R. Caseiro, P. Martins, J. Batista, "Exploiting the Circulant Structure of Tracking-by-detection with Kernels", ECCV 2012. [3] Y. Wu, J. Lim, M.-H. Yang, "Online Object Tracking: A Benchmark", CVPR 2013. Website: http://visual-tracking.net/ [4] P. Dollar, "Piotr's Image and Video Matlab Toolbox (PMT)". Website: http://vision.ucsd.edu/~pdollar/toolbox/doc/index.html [5] P. Dollar, S. Belongie, P. Perona, "The Fastest Pedestrian Detector in the West", BMVC 2010. _____________________________________ Copyright (c) 2014, Joao F. Henriques Permission to use, copy, modify, and distribute this software for research purposes with or without fee is hereby granted, provided that the above copyright notice and this permission notice appear in all copies. THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. 具有核化相关滤波器的高速跟踪 ?????????????J. F. Henriques R. Caseiro P. Martins J. Batista ???????????????????????????????????TPAMI 2014 ________________ 待出版。 arXiv预印：http：//arxiv.org/abs/1404.7584 项目网页：http：//www.isr.uc.pt/~henriques/circulant/ 这个MATLAB代码实现了一个基于Kernelized的简单跟踪管道 相关滤波器（KCF）和双相关滤波器（DCF）。 它可以免费用于研究。如果您发现它有用，请确认该文件 以上参考。 __________ 快速开始 1.在某处提取代码。 2.跟踪器准备在视觉跟踪的50个视频中的任何一个上运行 ???基准[3]。为此，它必须知道它们将位于/将位于何处。您可以 ???更改'download_videos.m'和'run_tracker.m'中的默认位置'base_path'。 3.如果您还没有视频，请运行'download_videos.m'（可能需要一些时间）。 4.执行不带参数的'run_tracker'以选择视频并在其上测试KCF。 注意：跟踪器使用Piotr工具箱中的'fhog'/'gradientMex'功能。 为方便起见，提供了一些预编译的MEX文件。如果他们不适合你 系统，只需从http://vision.ucsd.edu/~pdollar/toolbox/doc/index.html获取工具箱 __________ 主界面功能是'run_tracker'。您可以测试多种配置（KCF， DCF，MOSSE）通过不同的命令调用它： ?run_tracker ???没有任何参数，会要求您选择一个视频，跟踪使用 ???HOG上的高斯KCF，并以交互方式显示结果 ???数字。按'Esc'可以提前停止跟踪器。你可以导航 ???使用底部滚动条的视频。 ?run_tracker视频 ???允许您按名称选择视频。 'all'将运行所有视频 ???并显示平均统计数据。 'choose'将以交互方式选择一个。 ?run_tracker视频内核 ???选择一个内核。用于运行KCF的'高斯'/'多项式'，用于DCF的'线性'。 ?run_tracker视频内核功能 ???选择FEATURE类型，“hog”或“grey”（原始像素）。 ?run_tracker（VIDEO，KERNEL，FEATURE，SHOW_VISUALIZATION，SHOW_PLOTS） ???决定是否显示可滚动的图形和精确图形。 ?有用的组合： ?>> run_tracker选择高斯猪％Kernelized Correlation Filter（KCF） ?>> run_tracker选择线性猪％双相关滤波器（DCF） ?>> run_tracker选择高斯灰色％单通道KCF（ECCV'12纸） ?>> run_tracker选择线性灰色％MOSSE滤波器（单通道） 对于实际的跟踪代码，请查看“跟踪器”功能。 虽然不是必需的，但代码将使用MATLAB Parallel Computing 工具箱自动（如果可用）。 __________ 参考 [1] J. F. Henriques，R。Caseiro，P。Martins，J。Batista，“High-Speed Tracking with “核心相关滤波器”，TPAMI 2014（即将发布）。 [2] J. F. Henriques，R。Caseiro，P。Martins，J。Batista，“Exploiting the Circulant “使用内核进行检测的结构”，ECCV 2012。 [3] Y. Wu，J。Lim，M.-H。 Yang，“在线对象跟踪：基准”，CVPR 2013。 网站：http：//visual-tracking.net/ [4] P. Dollar，“Piotr的图像和视频Matlab工具箱（PMT）”。 网站：http：//vision.ucsd.edu/~pdollar/toolbox/doc/index.html [5] P. Dollar，S。Belongie，P。Perona，“中国最快的行人探测器” West“，BMVC 2010。 _____________________________________ 版权所有（c）2014，Joao F. Henriques 允许使用，复制，修改和分发此软件以进行研究 如有上述规定，特此授予有或没有费用的目的 版权声明和本许可声明出现在所有副本中。 软件按“原样”提供，作者不承担任何保证 关于本软件，包括所有暗示的保证 适销性和适应性。在任何情况下，作者都不应对此负责 任何特殊，直接，间接或间接损害或任何损害 无论是因为使用，数据还是利润的损失，都会导致这种情况发生 合同，疏忽或其他行为的行为，由此引起的 或与本软件的使用或性能有关。

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：