BioDrone Python Toolkit

UPDATE:
[2023.02.15] Update the toolkit installation, dataset download instructions and a concise example. Now the basic function of this toolkit has been finished.

This repository contains the official python toolkit for running experiments and evaluate performance on BioDrone. The code is written in pure python and is compile-free.

Table of Contents

• Python Toolkit
  • Table of Contents
    • Toolkit Installation
    • Dataset Download
    • A Concise Example
      • How to Define a Tracker?
      • How to Run Experiments?
      • How to Evaluate Performance?
    • Results of SOTA Trackers
    • Issues

Toolkit Installation

Clone the repository and install dependencies:

git clone https://github.com/huuuuusy/biodrone-toolkit-official.git
pip install -r requirements.txt

Then directly copy the biodrone folder to your workspace to use it.

Dataset Download

The BioDrone dataset includes 300 sequences, divided into three subsets (train/val/test).

The dataset download and file organization process is as follows：

• Download three subsets via download page in the project website (total: 142G).

• Check the number of files in each subset and run the unzipping script. Before unzipping:

  • the train subset should includ 300 files

  • the val subset should includ 100 files

  • the test subset should includ 200 files

• Run the unzipping script, and delete the script after decompression.

• Taking train subset as an example, the folder structure should follow:

|-- train/
|  |-- frame_002/
|  |  |-- 00000001.jpg/
|  |      ......
|  |  |-- 00000706.jpg/
|  |-- frame_004/
|  |   ......
|  |-- frame_597/
|  |-- frame_599/

• Unzip attribute.zip:

unzip attribute.zip -d attribute

• Attention that we only provide properties files for train and val subsets. For ground-truth files, we only offer the annotation of the first frame to initialize the model. You should upload final results for evaluation.

• Rename and organize folders as follows：

|-- BioDrone/
|  |-- data/
|  |  |-- train/
|  |  |  |-- frame_002/
|  |  |  |   ......
|  |  |  |-- frame_599/
|  |  |-- val/
|  |  |  |-- frame_005/
|  |  |  |   ......
|  |  |  |-- frame_594/
|  |  |-- test/
|  |  |  |-- frame_001/
|  |  |  |   ......
|  |  |  |-- frame_600/
|  |-- attribute/
|  |  |-- absent/
|  |  |-- blur_bbox/
|  |  |   ......
|  |  |-- restart/

A Concise Example

test.py is a simple example on how to use the toolkit to define a tracker, run experiments on dataset and evaluate performance.

How to Define a Tracker?

To define a tracker using the toolkit, simply inherit and override init and update methods from the Tracker class. You can find an example in this page. Here is a simple example:

from biodrone.trackers import Tracker

class IdentityTracker(Tracker):
    def __init__(self):
        super(IdentityTracker, self).__init__(
            name='IdentityTracker',  # tracker name
        )
    
    def init(self, image, box):
        self.box = box

    def update(self, image):
        return self.box

How to Run Experiments?

Instantiate an ExperimentBioDrone object, and leave all experiment pipelines to its run method:

from biodrone.experiments import ExperimentBioDrone

# ... tracker definition ...

# instantiate a tracker
tracker = IdentityTracker()

# setup experiment (validation subset)
experiment = ExperimentBioDrone(
  root_dir='SOT/BioDrone', # dataset's root directory
  save_dir= os.path.join(root_dir,'result'), # the path to save the experiment result
  subset='val', # 'train' | 'val' | 'test'
  repetition=1 
)
experiment.run(
  tracker, 
  visualize=False,
  save_img=False
  )

How to Evaluate Performance?

For evaluation in OPE mechanism, please use the report method of ExperimentBioDrone for this purpose:

# ... run experiments on dataset ...

# report tracking performance
experiment.report([tracker.name])

Results of SOTA Trackers on Testset

Metrics	OPE Mechanism	R-OPE Mechanism
precision plot
normalized precision plot
success plot

Issues

Please report any problems or suggessions in the Issues page.

Contributors

• Shiyu Hu
• Yipei Wang