ZebraZoom is open-source and can be downloaded and used free of charge: feel free to use it for your research and please acknowledge us in your publications.
Our software can be used for videos with a fixed background to track the heads and tails of freely swimming and of head-embedded zebrafish as well as the center of mass of any kind of animal.
View the user guide and the source code on Github.
Many software and methods have recently been developed to analyze animal behavior. Deciding which software/method to use for a specific project can often be difficult; and being able to use those software/methods efficiently can be even harder.
We first published an article on automated behavior analysis in 2013. Since then, we have been improving our software as well as testing and using software and methods created by others. We are also constantly following the latest developments in behavioral research, machine learning and computer vision.
As a paid service, we can thus help you analyze your behavioral experiments for any kind of animal. To do this, we will use open-source software developed by us and others. The choice of which software/method to use will be based exclusively on what's most relevant and cost effective for you.
Depending on the videos you want to analyze, we can help you use our tracking software ZebraZoom or the tracking software developed by other groups (DeepLabCut, IdTracker, Ctrax, etc...). These various software all have their strengths and weaknesses: we will choose which software to use depending on the videos you want to perform the tracking on. If none of these software correspond to what you need, we can even create new tracking software for you! Contact us to learn more.
Zebrafish larva tracked with ZebraZoom.
We have extensive experience analyzing the behavior of the zebrafish larvae, and we've used supervised and unsupervised machine learning to automatically classify bouts of movements into distinct behaviors. We can guide you through the most recent papers on automated behavior analysis, help you choose the most relevant method for your project, and even create new behavior analysis scripts for you! Contact us to learn more.
Unsupervised clustering of bouts.
In addition to using supervised and unsupervised machine learning to classify bouts, we've used deep learning to distinguish when a larva is rolling over. We've also used deep learning to track animals in videos. If you have an AI related project and need our assistance, feel free to contact us.
Rollover detection with ZZDeepRollover.
The software we've developed can easily be used through a graphical user interface, see the tutorial videos on how to use ZebraZoom. Similarly, while helping you on projects, we will make sure that all the code that we create or reuse from others is easily accessible to people with no coding experience.
ZebraZoom's graphical user interface.
Olivier is a software engineer and data scientist who has been working at the interface of biology and computer science for over a decade. In Virginia Tech he worked in a bioinformatics lab and participated in the creation of bio-security software for DNA sequence screening that he later presented along with his teammates at the iGEM competition and at the FBI headquarters. He has also worked in an early stage digital health startup in San Francisco creating a web-based chatbot relying on an AI medical diagnosis engine and has worked on other independent digital health projects. Olivier started working on ZebraZoom during his PhD in Claire Wyart’s lab in Paris and now wants to make other labs and corporations around the world benefit from his many years of experience analyzing the behavior of zebrafish larvae.
During her PhD, Claire developed novel methods for controlling the architecture of neuronal networks in vitro and demonstrated mechanisms underlying the emergence of spontaneous activity. She joined UC Berkeley for her postdoctoral fellowship and developed optogenetic methods in vivo by taking advantage of the transparency of the zebrafish larva. In this small vertebrate model, Claire used optogenetics to study sensory-motor integration, analyzing the processing of visual, mechanosensory and chemosensory pathways modulating locomotion. Since 2011, Claire started her team at the Brain and Spinal cord Institute in Paris. She is interested in deciphering the neuronal circuits that integrate information from the brain, from the periphery and from internal cues in order to modulate locomotion and posture as a function of inner physiological states.
We would like to thank Feng Quan, Laura Desban, Martin Carbó-Tano, Mathilde Lapoix, Mingyue Wu, Adna Dumitrescu, Jenna Sternberg, Maxime Kermarquer, Biswadeep Khan, Jane Yam, Claire Wyart's lab, Teresa Nicolson's lab and the Semmelhack Lab for letting us use the videos they've captured and for their help with the development of ZebraZoom. We would also like to thank Thomas Preat's lab and Nicolas Renier's lab for letting us use the videos they've captured.
In the unlikely event where the email address above wouldn't work (or if you email us and don't get any response), you can also contact us at: email@example.com
We welcome all questions and feedback regarding our open-source software ZebraZoom.
We are also available for hire if you need help analyzing animal behavior: see the paid services section above.
If you like our software, give us a star on Github and follow us on Twitter 😀.