pytracking-master.zip

# PyTracking A general python library for visual tracking algorithms. ## Table of Contents * [Running a tracker](#running-a-tracker) * [Overview](#overview) * [Trackers](#trackers) * [LWL](#LWL) * [KYS](#KYS) * [DiMP](#DiMP) * [ATOM](#ATOM) * [ECO](#ECO) * [Analysis](#analysis) * [Libs](#libs) * [Visdom](#visdom) * [VOT Integration](#vot-integration) * [Integrating a new tracker](#integrating-a-new-tracker) ## Running a tracker The installation script will automatically generate a local configuration file "evaluation/local.py". In case the file was not generated, run ```evaluation.environment.create_default_local_file()``` to generate it. Next, set the paths to the datasets you want to use for evaluations. You can also change the path to the networks folder, and the path to the results folder, if you do not want to use the default paths. If all the dependencies have been correctly installed, you are set to run the trackers. The toolkit provides many ways to run a tracker. **Run the tracker on webcam feed** This is done using the run_webcam script. The arguments are the name of the tracker, and the name of the parameter file. You can select the object to track by drawing a bounding box. **Note:** It is possible to select multiple targets to track! ```bash python run_webcam.py tracker_name parameter_name ``` **Run the tracker on some dataset sequence** This is done using the run_tracker script. ```bash python run_tracker.py tracker_name parameter_name --dataset_name dataset_name --sequence sequence --debug debug --threads threads ``` Here, the dataset_name is the name of the dataset used for evaluation, e.g. ```otb```. See [evaluation.datasets.py](evaluation/datasets.py) for the list of datasets which are supported. The sequence can either be an integer denoting the index of the sequence in the dataset, or the name of the sequence, e.g. ```'Soccer'```. The ```debug``` parameter can be used to control the level of debug visualizations. ```threads``` parameter can be used to run on multiple threads. **Run the tracker on a set of datasets** This is done using the run_experiment script. To use this, first you need to create an experiment setting file in ```pytracking/experiments```. See [myexperiments.py](experiments/myexperiments.py) for reference. ```bash python run_experiment.py experiment_module experiment_name --dataset_name dataset_name --sequence sequence --debug debug --threads threads ``` Here, ```experiment_module``` is the name of the experiment setting file, e.g. ```myexperiments``` , and ``` experiment_name``` is the name of the experiment setting, e.g. ``` atom_nfs_uav``` . **Run the tracker on a video file** This is done using the run_video script. ```bash python run_video.py tracker_name parameter_name videofile --optional_box optional_box --debug debug ``` Here, ```videofile``` is the path to the video file. You can either draw the box by hand or provide it directly in the ```optional_box``` argument. ## Overview The tookit consists of the following sub-modules. - [analysis](analysis): Contains scripts to analyse tracking performance, e.g. obtain success plots, compute AUC score. It also contains a [script](analysis/playback_results.py) to playback saved results for debugging. - [evaluation](evaluation): Contains the necessary scripts for running a tracker on a dataset. It also contains integration of a number of standard tracking and video object segmentation datasets, namely [OTB-100](http://cvlab.hanyang.ac.kr/tracker_benchmark/index.html), [NFS](http://ci2cv.net/nfs/index.html), [UAV123](https://ivul.kaust.edu.sa/Pages/pub-benchmark-simulator-uav.aspx), [Temple128](http://www.dabi.temple.edu/~hbling/data/TColor-128/TColor-128.html), [TrackingNet](https://tracking-net.org/), [GOT-10k](http://got-10k.aitestunion.com/), [LaSOT](https://cis.temple.edu/lasot/), [VOT](http://www.votchallenge.net), [Temple Color 128](http://www.dabi.temple.edu/~hbling/data/TColor-128/TColor-128.html), [DAVIS](https://davischallenge.org), and [YouTube-VOS](https://youtube-vos.org). - [experiments](experiments): The experiment setting files must be stored here, - [features](features): Contains tools for feature extraction, data augmentation and wrapping networks. - [libs](libs): Includes libraries for optimization, dcf, etc. - [notebooks](notebooks) Jupyter notebooks to analyze tracker performance. - [parameter](parameter): Contains the parameter settings for different trackers. - [tracker](tracker): Contains the implementations of different trackers. - [util_scripts](util_scripts): Some util scripts for e.g. generating packed results for evaluation on GOT-10k and TrackingNet evaluation servers, downloading pre-computed results. - [utils](utils): Some util functions. - [VOT](VOT): VOT Integration. ## Trackers The toolkit contains the implementation of the following trackers. ### LWL The official implementation for the LWL tracker ([paper](https://arxiv.org/pdf/2003.11540.pdf)). The tracker implementation file can be found at [tracker.lwl](tracker/lwl). ##### Parameter Files Two parameter settings are provided. These can be used to reproduce the results or as a starting point for your exploration. * **[lwl_ytvos](parameter/lwl/lwl_ytvos.py)**: The default parameter setting with ResNet-50 backbone which was used to generate YouTubeVOS results. * **[lwl_boxinit](parameter/lwl/lwl_boxinit.py)**: The parameters settings used to generate results with bounding box initialization on YouTubeVOS and DAVIS datasets. ### KYS The official implementation for the KYS tracker ([paper](https://arxiv.org/pdf/2003.11014.pdf)). The tracker implementation file can be found at [tracker.kys](tracker/kys). ##### Parameter Files * **[default](parameter/kys/default.py)**: The default parameter setting with ResNet-50 backbone which was used to produce all results in the paper, except on VOT and LaSOT. * **[default_vot](parameter/kys/default_vot.py)**: The parameters settings used to generate the VOT2018 results in the paper. ### DiMP The official implementation for the DiMP tracker ([paper](https://arxiv.org/abs/1904.07220)) and PrDiMP tracker ([paper](https://arxiv.org/abs/2003.12565)). The tracker implementation file can be found at [tracker.dimp](tracker/dimp). ##### Parameter Files * **[dimp18](parameter/dimp/dimp18.py)**: The default parameter setting with ResNet-18 backbone which was used to produce all DiMP-18 results in the paper, except on VOT. * **[dimp18_vot](parameter/dimp/dimp18_vot18.py)**: The parameters settings used to generate the DiMP-18 VOT2018 results in the paper. * **[dimp50](parameter/dimp/dimp50.py)**: The default parameter setting with ResNet-50 backbone which was used to produce all DiMP-50 results in the paper, except on VOT. * **[dimp50_vot](parameter/dimp/dimp50_vot18.py)**: The parameters settings used to generate the DiMP-50 VOT2018 results in the paper. * **[prdimp18](parameter/dimp/prdimp18.py)**: The default parameter setting with ResNet-18 backbone which was used to produce all PrDiMP-18 results in the paper, except on VOT. * **[prdimp50](parameter/dimp/prdimp50.py)**: The default parameter setting with ResNet-50 backbone which was used to produce all PrDiMP-50 results in the paper, except on VOT. * **[super_dimp](parameter/dimp/super_dimp.py)**: Combines the bounding-box regressor of PrDiMP with the standard DiMP classifier and better training and inference settings. The difference between the VOT and the non-VOT settings stems from the fact that the VOT protocol measures robustness in a very different manner compared to other benchmarks. In most benchmarks, it is highly important to be able to robustly *redetect* the target after e.g. an occlusion or brief target loss. On the other hand, in VOT the tracker is reset if the prediction does not overlap with the target on a *single* frame. This is then counted as a tracking