Latest Releases

Tracking Progress - Tracksys Director Dr Richard Lilley explores the development of behavioural measurement tools, from pen and paper to multi-modal software interfaces 29/12/2024

Software-based behaviour measurement tools for human-machine or human human interactions (and combinations of these) have been available for more than half a century. Initially of limited use for experienced practitioners and researchers, they’ve improved rapidly to become an important source of objective data. Many of these tools have their origin in university research ‘spinning out’ start-up companies that built up from modest beginnings to become large organisations with a wide portfolio of products.

Eyetrackers, for example, have moved on from the early days when they were large, cumbersome devices with limited spatial and temporal resolution that were often highly restrictive and uncomfortable for users. Now they’re sleek, lightweight devices that can be run using a small laptop or worn almost indistinguishably from conventional glasses. In addition, screen-based eyetrackers have added sophisticated stimulus presentation packages to show a mixture of images, videos and simulations, with interactive areas of interest that can react to where a user looks.

Mobile glasses-type systems have also evolved to allow easy data collection on areas of interest that are mapped to compensate for a mobile subject. These hardware advances are backed by powerful analysis tools matched with attractive visual outputs to serve the needs of publications and presentations respectively. Within ergonomics and human factors, there are a wide range of research areas for which eyetracking data is essential.

At Tracksys, we have clients who work across the rail, automotive and aviation sectors looking at a range of topics from health and safety to design and usability. For example, we recently supported a major client who was looking at control screens in a vehicle where eyetracking was vital in showing weaknesses in the design and also in an emergency reconfiguration, when one or more screens failed. Without knowing the gaze dwell-time, sequence and frequency of visits to key controls, it would have been impossible for the user to know the safest layout.

Product development in eyetracking has been driven by the demands of the market, but also by an ‘arms race’ between manufacturers trying to make bespoke products for niche markets or, conversely, increase the range of clients a flagship system can service. While this has been good for the research community, it’s also meant that periodically, small and medium size companies have been removed from the market. In the last ten years, even well-known brands such as SensoMotoric Instruments have disappeared when their intellectual property brought them to the attention of large global technology companies wanting to integrate eyetracking in their platforms.

That said, we’re now in a period of relative calm with incremental improvements to mature products, and a relatively stable user base. For those new to the technology, there’s also rich literature stretching back decades to draw on. In many ways, the development of eyetrackers has also ushered in a new era where the range of behavioural tools has rapidly expanded, especially where data can be collected automatically and without laborious post-processing.

If you’re beginning to add a behavioural dynamic to your research, you can now buy, for example, automatic emotion detection software which can be customised to fit your demographic, or physiological measurement devices that are quick and easy to fit to a subject, and provide far more data than the traditional heart rate measurement that comes with the cheapest sports watch.

Even manual scoring packages have increased in sophistication, backed by flexible design, easy coding and analysis capabilities that allow hypotheses to be tested within the interface. That said, it will probably come as no surprise that manufacturers have, by and large, stayed within the narrow ‘channel’ that their brand is well known for. This makes good business sense. The upside of this approach is rapid development and increasing capability, but this often comes at the cost of limited cross-connectivity with other devices, although there is some. This puts a frustrating limit, or a least an increased workload, on researchers who want to do multi modal data collection across a range of voluntary and involuntary behavioural responses. In practice, this means collecting data linked by an initial time-stamp for post-processing in, say Excel or MatLab, before export to an analysis tool in a prescribed format.

It’s been a long sought-after goal to create a unifying platform for multiple behavioural measures to flow into for easy acquisition and analysis (with the added bonus of hypothesis testing using one or more data streams to probe the rest of the data). Such an endeavour comes with significant challenges, for example: 

• How quickly and easily can a subject be prepared for data collection?

• Can all data be streamed seamlessly from the devices used?

• Can different sample rates from different devices be accommodated?

Even manual scoring packages have increased in sophistication, backed by flexible design, easy coding and analysis capabilities that allow hypotheses to be tested within the interface. That said, it will probably come as no surprise that manufacturers have, by and large, stayed within the narrow ‘channel’ that their brand is well known for. This makes good business sense. The upside of this approach is rapid development and increasing capability, but this often comes at the cost of limited cross-connectivity with other devices, although there is some. This puts a frustrating limit, or a least an increased workload, on researchers who want to do multi modal data collection across a range of voluntary and involuntary behavioural responses. In practice, this means collecting data linked by an initial time

• How can ‘drift’ in data streams be controlled because a simple trigger to synchronise data when collection starts is not enough?

• Is it possible to make a system that can also present stimuli flexibly, collect and also analyse data?

• Is this a viable product that meets user needs and a business case?

The last of these questions has been, I think, the main stumbling block. No one would deny that the next logical step after creating behavioural tools that measure specific things (like where you look, or how you feel) is to have everything ‘under one roof’. I’ve touched on the issue already though – brands seek to cement their reputations as providers of, say, ‘the world’s fastest eyetracker’ or the ‘least invasive physiological monitor’.

Fortunately, 2024 has finally seen the arrival of an easy-to-use fully integrated system in the shape of NoldusHub which addresses all of the points raised above. This is not to say that ergonomics and human factors researchers may not still selectively choose from the behavioural measurement tools available, but it does finally give them a complete solution in one easy-to-use package.

Tracksys will present a Masterclass on behavioural measurement tools and their integration at Ergonomics & Human Factors 2025.

This article first appeared in the December 2024 issue of The Ergonomist, the magazine of the Chartered Institute of Ergonomics and Human Factors

Bertec force measurement products 20/12/2022

Full range of force measurement devices from Bertec available.

• Instrumented treadmill
• Immersive lab
• Force plates - portable and fixed plates
• Portable staircase - instrumented stairs
• Instrumented Handrails 

EthoVision XT 16 Released 01/06/2021

EthoVision XT 16 marks the start of the 'Deep Learning' era using neural networks to allow a computer to ‘learn by experience’. This drastically improves the detection of body points in rats and mice. The result: unsurpassed video tracking quality!

The deep learning-based detection of the nose point, center, and tail base of rats and mice truly takes video tracking to a new level, providing more accurate and stable body points detection. This is especially great for social preference testing, novel object recognition tests, hole board tests, elevated plus maze, etc.

Tobii Pro Glasses 3 03/06/2020

  New wearable eye tracking glasses from Tobii Pro:

• Easy to wear in real world scenarios
• Embedded 4 cameras for accurate gaze data
• Wide-angled scene camera capturing greater field of view
• Slippage compensation giving robust data
• Add-on IR blocking, safety lenses for outdoors and industrial environments
• Snap-on corrective lenses
• Synchronise with physiological data

FaceReader 8.1 26/05/2020

 Updated version 8 offers a number of improvements:

• New reporting display – allows users to select analysis windows of interest, in their preferred layout

• Improved batch analysis – analysed data saves automatically as it’s processed

• Improved Experimental Consumption Behaviour Module includes chewing, biting and sipping, and is still free to use

Tobii Pro Fusion 17/10/2019

New screen-based eye tracker from Tobii Pro:

• Up to 250Hz sampling rate for more detailed data
• Fully portable with no external processing unit
• Two eye cameras and improved algorithms giving greater head movement tolerance
• Pupil size data reported
• Eye images provided

EthoVision 15 released 09/09/2019

EthoVision XT 15 is here, with all these new features:

• Body Point Detection – new intuitive names with improvements in the “Any species/default” method and Nose-Tail detection
• Trial list – when clearing a trial to re-do the tracking, the link to the video is now maintained
• Behaviour recognition – Subject Exposure Indicator added and flexibility on body point detection method now available
• Data analysis – heatmap scale added to help quantify time of animal spent in a location
• Data management – Backups can now be made without the dongle being inserted
• Video tutorials – subtitles added and new tutorial on Behaviour Recognition Module
• External Data – new import profile for Ponemah files from Data Sciences International
• Basler settings can now be changed within EthoVision XT, instead of Pylon
• 2 webcams now supported
• Up to 16 IP cameras can be tracked

Viso 8 25/03/2019

New features to the video and audio recording software are:

•     Viso Web App – a web based application for remote viewing and controlling recordings from any device
•     Audit trail – allows supervision of all user actions to help comply with legislation
•     Pre-roll interval – define the delay to start playing back the video before a marked event actually starts

FaceReader 8 and Baby FaceReader released 11/01/2019

Here are the latest features and modules found in the newest version of FaceReader:

•     Baby FaceReader licence – facial expressions in infants ageing 6 to 24 months
•     Measure eating behaviour
•     Audio recording and playback with video
•     Analysis in low light conditions
•     Action units now analysed for both left and right sides of the face
•     Design your own expressions within the Action Unit Module
•     Heart rate variability now available within the RPPG Module

EthoVision XT 14 released 09/07/2018

New to the EthoVision XT software:

•          Statistics layout – customise to your needs
•          USB and IP cameras now supported
•          Analysis of minimum and maximum duration of state parameters
•          Copy and paste blocks in trial control and data selection