Feeling very privileged to be a part of lots of exciting demonstrations and discussions of cool work at #CHI2024. From the PhysioCHI workshop, to the REVEAL interactivity, and our best paper honourable mention talk🏅. Happy to have met so many great folks! 😁

And a big thankyou to the all the folks at the Southwest Pre-CHI event for listening to a preview of my talk and voting for me as a powerpoint prodigy! What a success the event was, will definitely be in attendance next year! @UWEBristol @GW4Alliance 14/12

If you are interested in this work come to the paper session: programs.sigchi.org/chi/2024… And if you would like to try the VR exergames and other exciting VR projects, please find the REVEAL interactivity during the demo sessions: programs.sigchi.org/chi/2024… 12/12

Theres lots of avenues for future work. We hope to help exergame designers and emotion recognition researchers get started with our available artifacts: Extended analysis report, aggregate + realtime dataset (n=72), & the VEs/software used to collect & clean the data. 📊🖥️11/12

We have lots of recommendations for doing emotion recognition in VR Exergames, here is a quick summary: 10/12

There is so much to discuss with this work and the emotion estimation models do not just have applicability in VR Exergames, but any context that involves physical activity! (e.g. sports and health, physical rehabilitation, and physically demanding work). 9/12

We used different data cleaning techniques on physiological data. These techniques really impact the utility of certain measures. For example, SCL is often used to detect arousal, but during exercise if you dont clean the measure it actually makes emotion estimation worse! 8/12

We then modelled the relationship of different emotions to 10 physiological metrics, producing regression models that could estimate emotion. 📈 The table shows on overview of significant measures for each type of emotion. 6/12

We first investigated if we could elicit separate emotions reliably during VR exergaming covering a wide range of emotions to help produce our emotion models. 😄😟😬😌 We designed 4 different exergame experiences based on surveying lots of literature on emotion elicitation. 5/12

So we decided to see if we could still estimate emotions using physiology even while people engage in intense exercise 🚴‍♀️🥵 One of the most challenging contexts for emotion recognition! 4/12