DDES9902 Chap.4 Biometrics, Simulation and Visualisation
Biometrics, Simulation and Visualisation
Tracked gaze, facial movement and body pose can extend immersive interaction, but the signal a device records is not identical to the intention, attention or emotion a designer assigns to it. This chapter separates sensed behaviour, interpreted state and adaptive response. It also treats visualisation and simulation as representations with a task and a boundary.
Immersion is useful when it makes a relationship easier to inspect or an action possible to rehearse. The report must state what was measured, transformed or omitted and limit decisions when the interpretation or model is uncertain. A defensible biometric feature begins with calibration and a stated sampling condition.
Designers should record signal loss, latency, occlusion and the behaviour used to confirm or reject an interpretation. A gaze dwell may support selection without proving attention, while facial movement may animate an avatar without authorising an emotion score. The permitted response should become less consequential as uncertainty increases, and users need a way to correct or disable an inference.
Data handling is part of the interaction: collection purpose, retention, access and deletion shape whether consent remains meaningful. For visualisation, compare the immersive version with a simpler screen or physical representation on the same task. Measure comprehension, navigation or decision quality rather than assuming three-dimensional presentation is superior.
For simulation, identify which model relationships are reliable, which variables are simplified and what a participant could wrongly generalise. These controls turn tracking and immersion into bounded design tools rather than claims that a system can transparently read a person or reproduce reality.
What this chapter covers
- 01
Eye, face and body tracking as sources of interaction signals
- 02
Difference between measured behaviour and inferred user state
- 03
Calibration, occlusion and individual variability
- 04
Foveated rendering and gaze-assisted interaction
- 05
Visualisation chosen for a defined analytical task
- 06
Simulation boundaries, omissions and appropriate claims
Worked example · free
Use gaze without treating every look as a choice
- +1Use gaze to reveal a preview after a short dwell rather than immediately selecting the item.
- +1Require a separate confirmation input so looking is not equated with choosing.
- +1Show dwell progress, allow cancellation and provide a ray-pointer alternative.
- +1Record false previews and failed confirmations while avoiding storage of an attention score.
Key terms
- Gaze tracking
- Measurement of eye orientation or fixation used to estimate where a person is looking.
- Foveated rendering
- A rendering method that concentrates detail near the gaze region and reduces work elsewhere.
- Calibration
- The process of aligning a sensor's output with an individual user and the current setup.
- Biometric inference
- An interpretation about a person's state drawn from measured bodily or behavioural signals.
- Simulation
- A designed representation of a system or situation used to explore behaviour or rehearse action.
- Data visualisation
- A representation that makes selected patterns, quantities or relationships available for inspection.
Biometrics, Simulation and Visualisation FAQ
Can gaze tracking tell whether someone is interested?
It can show where the system estimates a person looked, under calibration and sampling limits. Interest is an interpretation that may also depend on confusion, task demands or distraction. Use gaze for a narrow interaction or research purpose and avoid treating it as a direct reading of mental state.
When does immersive visualisation add value?
It helps when spatial scale, depth, relation or embodied inspection supports the analytical task. If a flat chart enables faster and more accurate comparison, adding three-dimensional form can create occlusion and navigation cost. The representation should be judged by what users can understand or decide.
How should a simulation limitation be reported?
Name the omitted or simplified behaviour and explain which decisions remain appropriate. For example, a trainer may reproduce action order while simplifying forces. The limitation belongs beside the conclusion because visual credibility can otherwise invite users to generalise beyond the model.
Assessment move
For each tracked input, make a three-column table: measured signal, possible interpretations and permitted system actions. Add at least one alternative explanation and one accessible fallback. This prevents the common leap from bodily data to a confident claim about intention or emotion. Compare representations by the task they enable: lookup, comparison, navigation, rehearsal or communication.
Remove visual dimensions that do not encode a meaningful relationship. When reviewing a simulation, list what is physically represented, what behaves like the target system and what psychological demand is reproduced. Then write a boundary sentence about what the prototype cannot establish.
Apply the same exercise to your project and include the chosen authority limit in the test plan, so evaluation checks false actions and recovery as well as successful interaction.
