MDNet-master
所属分类:matlab编程
开发工具:matlab
文件大小:147KB
下载次数:46
上传日期:2017-01-10 20:49:49
上 传 者:
vvruir
说明: 一个基于CNN的目标跟踪程序,效果非常好。成功率89
(Target tracking based on a CNN program, the effect is very good)
文件列表:
compile_matconvnet.m (258, 2016-07-12)
LICENSE (2514, 2016-07-12)
matconvnet\COPYING (735, 2016-07-12)
matconvnet\Makefile (7759, 2016-07-12)
matconvnet\Makefile.mex (793, 2016-07-12)
matconvnet\Makefile.nvcc (925, 2016-07-12)
matconvnet\matlab\src\bits\data.cpp (106, 2016-07-12)
matconvnet\matlab\src\bits\data.cu (9625, 2016-07-12)
matconvnet\matlab\src\bits\data.hpp (4239, 2016-07-12)
matconvnet\matlab\src\bits\datacu.cu (6171, 2016-07-12)
matconvnet\matlab\src\bits\datacu.hpp (2237, 2016-07-12)
matconvnet\matlab\src\bits\datamex.cpp (109, 2016-07-12)
matconvnet\matlab\src\bits\datamex.cu (11354, 2016-07-12)
matconvnet\matlab\src\bits\datamex.hpp (1604, 2016-07-12)
matconvnet\matlab\src\bits\impl\blashelper.hpp (5132, 2016-07-12)
matconvnet\matlab\src\bits\impl\copy.hpp (874, 2016-07-12)
matconvnet\matlab\src\bits\impl\copy_cpu.cpp (595, 2016-07-12)
matconvnet\matlab\src\bits\impl\copy_gpu.cu (625, 2016-07-12)
matconvnet\matlab\src\bits\impl\fast_mutex.h (6940, 2016-07-12)
matconvnet\matlab\src\bits\impl\im2row.hpp (2964, 2016-07-12)
matconvnet\matlab\src\bits\impl\im2row_cpu.cpp (7752, 2016-07-12)
matconvnet\matlab\src\bits\impl\im2row_gpu.cu (10389, 2016-07-12)
matconvnet\matlab\src\bits\impl\imread_gdiplus.cpp (4666, 2016-07-12)
matconvnet\matlab\src\bits\impl\imread_helpers.hpp (14063, 2016-07-12)
matconvnet\matlab\src\bits\impl\imread_libjpeg.cpp (5943, 2016-07-12)
matconvnet\matlab\src\bits\impl\imread_quartz.cpp (5267, 2016-07-12)
matconvnet\matlab\src\bits\impl\nnconv_blas.hpp (12684, 2016-07-12)
matconvnet\matlab\src\bits\impl\nnconv_cudnn.cu (15350, 2016-07-12)
matconvnet\matlab\src\bits\impl\nnconv_cudnn.hpp (2314, 2016-07-12)
matconvnet\matlab\src\bits\impl\nnpooling_cudnn.cu (8135, 2016-07-12)
matconvnet\matlab\src\bits\impl\nnpooling_cudnn.hpp (2549, 2016-07-12)
matconvnet\matlab\src\bits\impl\normalize.hpp (2769, 2016-07-12)
matconvnet\matlab\src\bits\impl\normalize_cpu.cpp (12401, 2016-07-12)
matconvnet\matlab\src\bits\impl\normalize_gpu.cu (5132, 2016-07-12)
matconvnet\matlab\src\bits\impl\pooling.hpp (6307, 2016-07-12)
matconvnet\matlab\src\bits\impl\pooling_cpu.cpp (10332, 2016-07-12)
matconvnet\matlab\src\bits\impl\pooling_gpu.cu (12545, 2016-07-12)
... ...
## MDNet: Multi-Domain Convolutional Neural Network Tracker
Created by [Hyeonseob Nam](https://kr.linkedin.com/in/hyeonseob-nam/) and [Bohyung Han](http://cvlab.postech.ac.kr/~bhhan/) at POSTECH
Project Webpage: http://cvlab.postech.ac.kr/research/mdnet/
### Introduction
MDNet is the state-of-the-art visual tracker based on a CNN trained on a large set of tracking sequences,
and the winner tracker of [The VOT2015 Challenge](http://www.votchallenge.net/vot2015/).
Detailed description of the system is provided by our [arXiv technical report](http://arxiv.org/abs/1510.07945).
This software is implemented using [MatConvNet](http://www.vlfeat.org/matconvnet/) and part of [R-CNN](https://github.com/rbgirshick/rcnn).
### Citation
If you're using this code in a publication, please cite our paper.
@InProceedings{nam2016mdnet,
author = {Nam, Hyeonseob and Han, Bohyung},
title = {Learning Multi-Domain Convolutional Neural Networks for Visual Tracking},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2016}
}
### License
This software is being made available for research purpose only.
Check LICENSE file for details.
### System Requirements
This code is tested on *** bit Linux (Ubuntu 14.04 LTS).
**Prerequisites**
0. MATLAB (tested with R2014a)
0. MatConvNet (tested with version 1.0-beta10, included in this repository)
0. For GPU support, a GPU (~2GB memory) and CUDA toolkit according to the [MatConvNet installation guideline](http://www.vlfeat.org/matconvnet/install/) will be needed.
### Installation
0. Compile MatConvNet according to the [installation guideline](http://www.vlfeat.org/matconvnet/install/). An example script is provided in 'compile_matconvnet.m'.
0. Run 'setup_mdnet.m' to set the environment for running MDNet.
### Online Tracking using MDNet
**Pretrained Models**
If you only need to run the tracker, you can use the pretrained MDNet models:
0. models/mdnet_vot-otb.mat (trained on VOT13,14,15 excluding OTB)
0. models/mdnet_otb-vot14.mat (trained on OTB excluding VOT14)
0. models/mdnet_otb-vot15.mat (trained on OTB excluding VOT15)
**Demo**
0. Run 'tracking/demo_tracking.m'.
The demo performs online tracking on *'Diving'* sequence using a pretrained model 'models/mdnet_vot-otb.mat'.
In case of out of GPU memory, decrease *opts.batchSize_test* in 'tracking/mdnet_init.m'.
You can also disable the GPU support by setting *opts.useGpu* in 'tracking/mdnet_init.m' to false (not recommended).
### Learning MDNet
**Preparing Datasets**
You may need OTB and VOT datasets for learning MDNet models. You can also use other datasets by configuring 'utils/genConfig.m'.
0. Download [OTB](http://cvlab.hanyang.ac.kr/tracker_benchmark/datasets.html) and [VOT](http://www.votchallenge.net/) datasets.
0. Locate the OTB sequences in 'dataset/OTB' and VOT201x sequences in 'dataset/VOT/201x', or modify the variables *benchmarkSeqHome* in 'utils/genConfig.m' properly.
**Demo**
0. Run 'pretraining/demo_pretraining.m'.
The demo trains new MDNet models using OTB or VOT sequences.
近期下载者:
相关文件:
收藏者: