Python package for multiple object tracking research with focus on laboratory animals tracking.

Related tags

Deep Learningmotutils
Overview

Build Status

motutils is a Python package for multiple object tracking research with focus on laboratory animals tracking.

Features

  • loads:
  • saves: MOTChallenge CSV
  • Mot, BboxMot and PoseMot classes backed by xarray dataset with frame and id coordinates
  • export to Pandas DataFrame
  • oracle detector: fake all knowing detector based on ground truth with configurable inaccuracies
  • different classes of tracked objects: point, bounding box, pose
  • interpolation of missing positions
  • find mapping between MOT results and ground truth
  • visualization:
    • tracked positions / objects overlaid on a video
    • montage of multiple videos with results and/or ground truth
  • cli
    • visualization
    • evaluation ()
    • mot format conversion

visualization montage

Video comparison of multiple tracking methods and the ground truth.

Installation

pip install git+https://github.com/smidm/motutils

Usage

$ motutils --help
Usage: motutils [OPTIONS] COMMAND [ARGS]...

Options:
--load-mot FILENAME             load a MOT challenge csv file(s)
--load-gt FILENAME              load ground truth from a MOT challenge csv
file
--load-idtracker FILENAME       load IdTracker trajectories (e.g.,
trajectories.txt)
--load-idtrackerai FILENAME     load idtracker.ai trajectories (e.g.,
trajectories_wo_gaps.npy)
--load-sleap-analysis FILENAME  load SLEAP analysis trajectories (exported
from sleap-label File -> Export Analysis
HDF5)
--load-toxtrac FILENAME         load ToxTracker trajectories (e.g.,
Tracking_0.txt)
--toxtrac-topleft-xy 
   
    ...
position of the arena top left corner, see
first tuple in the Arena line in Stats_1.txt
--help                          Show this message and exit.

Commands:
convert    Convert any format to MOT Challenge format.
eval       Evaluate a single MOT file against the ground truth.
visualize  Visualize MOT file(s) overlaid on a video.

   
$ motutils convert --help

Usage: motutils convert [OPTIONS] OUTPUT_MOT

  Convert any format to MOT Challenge format.


$ motutils eval --help

Usage: motutils eval [OPTIONS]

  Evaluate a single MOT file against the ground truth.

Options:
  --write-eval FILENAME  write evaluation results as a CSV file
  --keypoint INTEGER     keypoint to use when evaluating pose MOT results
                         against point ground truth

$ motutils visualize --help

Usage: motutils visualize [OPTIONS] VIDEO_IN VIDEO_OUT
                          [SOURCE_DISPLAY_NAME]...

  Visualize MOT file(s) overlaid on a video.

Options:
  --limit-duration INTEGER  visualization duration limit in s
  --help                    Show this message and exit.

Python API Quickstart

>> mot.ds Dimensions: (frame: 4500, id: 5) Coordinates: * frame (frame) int64 0 1 2 3 4 5 6 ... 4494 4495 4496 4497 4498 4499 * id (id) int64 1 2 3 4 5 Data variables: x (frame, id) float64 434.5 277.7 179.2 ... 185.3 138.6 420.2 y (frame, id) float64 279.0 293.6 407.9 ... 393.3 387.2 294.7 width (frame, id) float64 nan nan nan nan nan ... nan nan nan nan nan height (frame, id) float64 nan nan nan nan nan ... nan nan nan nan nan confidence (frame, id) float64 1.0 1.0 1.0 1.0 1.0 ... 1.0 1.0 1.0 1.0 1.0 >>> mot.num_ids() 5 >>> mot.count_missing() 0 >>> mot.get_object(frame=1, obj_id=2) Dimensions: () Coordinates: frame int64 1 id int64 2 Data variables: x float64 278.2 y float64 293.7 width float64 nan height float64 nan confidence float64 1.0 >>> mot.match_xy(frame=1, xy=(300, 300), maximal_match_distance=40) Dimensions: () Coordinates: frame int64 1 id int64 2 Data variables: x float64 278.2 y float64 293.7 width float64 nan height float64 nan confidence float64 1.0 >>> mot.to_dataframe() frame id x y width height confidence 0 1 1 434.5 279.0 -1.0 -1.0 1.0 1 1 2 277.7 293.6 -1.0 -1.0 1.0 2 1 3 179.2 407.9 -1.0 -1.0 1.0 3 1 4 180.0 430.0 -1.0 -1.0 1.0 4 1 5 155.0 397.0 -1.0 -1.0 1.0 ... .. ... ... ... ... ... 22495 4500 1 90.3 341.9 -1.0 -1.0 1.0 22496 4500 2 187.9 431.9 -1.0 -1.0 1.0 22497 4500 3 185.3 393.3 -1.0 -1.0 1.0 22498 4500 4 138.6 387.2 -1.0 -1.0 1.0 22499 4500 5 420.2 294.7 -1.0 -1.0 1.0 [22500 rows x 7 columns]"> 
>>> from motutils import Mot
>>> mot = Mot("tests/data/Sowbug3_cut.csv")

>>> mot.ds
<xarray.Dataset>
Dimensions:     (frame: 4500, id: 5)
Coordinates:
  * frame       (frame) int64 0 1 2 3 4 5 6 ... 4494 4495 4496 4497 4498 4499
  * id          (id) int64 1 2 3 4 5
Data variables:
    x           (frame, id) float64 434.5 277.7 179.2 ... 185.3 138.6 420.2
    y           (frame, id) float64 279.0 293.6 407.9 ... 393.3 387.2 294.7
    width       (frame, id) float64 nan nan nan nan nan ... nan nan nan nan nan
    height      (frame, id) float64 nan nan nan nan nan ... nan nan nan nan nan
    confidence  (frame, id) float64 1.0 1.0 1.0 1.0 1.0 ... 1.0 1.0 1.0 1.0 1.0

>>> mot.num_ids()
5

>>> mot.count_missing()
0

>>> mot.get_object(frame=1, obj_id=2)
<xarray.Dataset>
Dimensions:     ()
Coordinates:
    frame       int64 1
    id          int64 2
Data variables:
    x           float64 278.2
    y           float64 293.7
    width       float64 nan
    height      float64 nan
    confidence  float64 1.0

>>> mot.match_xy(frame=1, xy=(300, 300), maximal_match_distance=40)
<xarray.Dataset>
Dimensions:     ()
Coordinates:
    frame       int64 1
    id          int64 2
Data variables:
    x           float64 278.2
    y           float64 293.7
    width       float64 nan
    height      float64 nan
    confidence  float64 1.0

>>> mot.to_dataframe()
       frame  id      x      y  width  height  confidence
0          1   1  434.5  279.0   -1.0    -1.0         1.0
1          1   2  277.7  293.6   -1.0    -1.0         1.0
2          1   3  179.2  407.9   -1.0    -1.0         1.0
3          1   4  180.0  430.0   -1.0    -1.0         1.0
4          1   5  155.0  397.0   -1.0    -1.0         1.0
      ...  ..    ...    ...    ...     ...         ...
22495   4500   1   90.3  341.9   -1.0    -1.0         1.0
22496   4500   2  187.9  431.9   -1.0    -1.0         1.0
22497   4500   3  185.3  393.3   -1.0    -1.0         1.0
22498   4500   4  138.6  387.2   -1.0    -1.0         1.0
22499   4500   5  420.2  294.7   -1.0    -1.0         1.0
[22500 rows x 7 columns]

Documentation

See the quickstart and tests for now.

Write me if you would like to use the package, but the lack of documentation is hindering you. You can easily reorder my priorities on this simply just by letting me know that there is an interest.

Owner
Matěj Šmíd
Matěj Šmíd
GitHub Repository
DvD-TD3: Diversity via Determinants for TD3 version

DvD-TD3: Diversity via Determinants for TD3 version The implementation of paper Effective Diversity in Population Based Reinforcement Learning. Instal

3 Feb 11, 2022
A framework that constructs deep neural networks, autoencoders, logistic regressors, and linear networks

A framework that constructs deep neural networks, autoencoders, logistic regressors, and linear networks without the use of any outside machine learning libraries - all from scratch.

Kordel K. France 2 Nov 14, 2022
Reinforcement Learning for Automated Trading

Reinforcement Learning for Automated Trading This thesis has been realized for the obtention of the Master's in Mathematical Engineering at the Polite

Pierpaolo Necchi 80 Jun 19, 2022
Classify music genre from a 10 second sound stream using a Neural Network.

MusicGenreClassification Academic research in the field of Deep Learning (Deep Neural Networks) and Sound Processing, Tel Aviv University. Featured in

Matan Lachmish 453 Dec 27, 2022
Securetar - A streaming wrapper around python tarfile and allow secure handling files and support encryption

Secure Tar Secure Tarfile library It's a streaming wrapper around python tarfile

Pascal Vizeli 2 Dec 09, 2022
A mini library for Policy Gradients with Parameter-based Exploration, with reference implementation of the ClipUp optimizer from NNAISENSE.

PGPElib A mini library for Policy Gradients with Parameter-based Exploration [1] and friends. This library serves as a clean re-implementation of the

NNAISENSE 56 Jan 01, 2023
A PyTorch implementation of SlowFast based on ICCV 2019 paper "SlowFast Networks for Video Recognition"

SlowFast A PyTorch implementation of SlowFast based on ICCV 2019 paper SlowFast Networks for Video Recognition. Requirements Anaconda PyTorch conda in

Hao Ren 8 Dec 23, 2022
A simple approach to emable dense segmentation with ViT.

Vision Transformer Segmentation Network This implementation of ViT in pytorch uses a super simple and straight-forward way of generating an output of

HReynaud 5 Jan 03, 2023
NLP From Scratch Without Large-Scale Pretraining: A Simple and Efficient Framework

NLP From Scratch Without Large-Scale Pretraining This repository contains the code, pre-trained model checkpoints and curated datasets for our paper:

Xingcheng Yao 224 Dec 08, 2022
TorchX: A PyTorch Extension Library for More Efficient Deep Learning

TorchX TorchX: A PyTorch Extension Library for More Efficient Deep Learning. @misc{torchx, author = {Ansheng You and Changxu Wang}, title = {T

Donny You 8 May 28, 2022
This repository contains the source code of our work on designing efficient CNNs for computer vision

Efficient networks for Computer Vision This repo contains source code of our work on designing efficient networks for different computer vision tasks:

Sachin Mehta 386 Nov 26, 2022
Ranger - a synergistic optimizer using RAdam (Rectified Adam), Gradient Centralization and LookAhead in one codebase

Ranger-Deep-Learning-Optimizer Ranger - a synergistic optimizer combining RAdam (Rectified Adam) and LookAhead, and now GC (gradient centralization) i

Less Wright 1.1k Dec 21, 2022
Cold Brew: Distilling Graph Node Representations with Incomplete or Missing Neighborhoods

Cold Brew: Distilling Graph Node Representations with Incomplete or Missing Neighborhoods Introduction Graph Neural Networks (GNNs) have demonstrated

37 Dec 15, 2022
The source code and dataset for the RecGURU paper (WSDM 2022)

RecGURU About The Project Source code and baselines for the RecGURU paper "RecGURU: Adversarial Learning of Generalized User Representations for Cross

Chenglin Li 17 Jan 07, 2023
Pytorch Implementation of Auto-Compressing Subset Pruning for Semantic Image Segmentation

Pytorch Implementation of Auto-Compressing Subset Pruning for Semantic Image Segmentation Introduction ACoSP is an online pruning algorithm that compr

Merantix 8 Dec 07, 2022
An easier way to build neural search on the cloud

An easier way to build neural search on the cloud Jina is a deep learning-powered search framework for building cross-/multi-modal search systems (e.g

Jina AI 17k Jan 02, 2023
Realtime_Multi-Person_Pose_Estimation

Introduction Multi Person PoseEstimation By PyTorch Results Require Pytorch Installation git submodule init && git submodule update Demo Download conv

tensorboy 1.3k Jan 05, 2023
Official PyTorch implementation of Less is More: Pay Less Attention in Vision Transformers.

Less is More: Pay Less Attention in Vision Transformers Official PyTorch implementation of Less is More: Pay Less Attention in Vision Transformers. By

73 Jan 01, 2023
Google Brain - Ventilator Pressure Prediction

Google Brain - Ventilator Pressure Prediction https://www.kaggle.com/c/ventilator-pressure-prediction The ventilator data used in this competition was

Samuele Cucchi 1 Feb 11, 2022
Vector Quantized Diffusion Model for Text-to-Image Synthesis

Vector Quantized Diffusion Model for Text-to-Image Synthesis Due to company policy, I have to set microsoft/VQ-Diffusion to private for now, so I prov

Shuyang Gu 294 Jan 05, 2023