Work Package 13 (ER 3)

Developing novel automated rhythmic behaviour analyses

General details

UKLocation
Actual Analytics (UK)

Supervisor
Prof J Douglas Armstrong

ER3
TBA

Other partners involved

Secondments

The secondments depend on the individual training requirements of the ER. Preferentially,
the ER will learn the specialist requirements for Nasonia courtship at Groningen and the requirements for rhythmic behaviour analysis at the other partner sites.

Objectives

  • Develop new rhythmic behaviour analysis systems within the INsecTime consortium and beyond.
  • Develop, deploy and evaluate systems that meet requested requirements

Methods

Develop, test and validate the various novel algorithms and hardware required

Description

Developing novel automated rhythmic behaviour analyses
​Breakthroughs in recent years in algorithms that both track and recognise behaviour from video means that not only can gross locomotion be detected, but more subtle behaviours can be captured automatically. Actual Analytics specialises in the management and handling of video based behaviour data and now markets the first on-line behavioural analysis system that allows data warehousing and sharing. Actual also develops advanced algorithms for tracking and behaviour recognition. The current model requires behaviour researchers to upload video files to an on-line service (based on Amazon) where high performance computing systems can apply the latest algorithms in tracking and analysis to extract the highest quality data. This model works extremely well with short experiments. Actual can use third party video capture software to collect the data but has also designed its own client-side system (iComply) that helps non-specialist users optimise video capture (a non-trivial process). iComply is currently implemented to support short duration experiments. The computer scientist ER3 will work with our development team and in collaboration with the behaviour groups in the consortium to extend this software to enable long-term video capture and analysis of rhythmic behaviour, whether short-term fluctuations in Nasonia courtship (in collaboration with Groningen group), or circadian 24 h rhythms (most of the other groups).  The programme of work will require the development of systems that can handle changes in lighting. We also want to implement a system that can automatically detect the start and end of a trial so that the camera may simply be left live. We will also build in support for environment control by the software itself so the system can be left in a closed loop with minimal researcher intervention.  The ER will also help provide support to our collaborators in terms of choosing cameras, lighting and experimental configuration while themselves learning about and contributing to the analysis of the behavioural phenomena under investigation.

Results and milestones

  1. Month 6 Requirements analysis for automated rhythmic behaviour analysis completed
  2. Month 18 Development of 24h tracking version of Actual software
  3. Month 30 Application of video tracking to clock behaviours

Synergies, Risks & Exploitation

Actual will enhance all behavioural projects that require long-term behavioural monitoring. Risks, minimal. Actual will incorporate this work into their product line to be marketed internationally.

See more Work Packages in Research Area C