... | ... | @@ -79,6 +79,26 @@ Overall there must be a great deal of user control over the configuration of the |
|
|
## 2.1 USER NEEDS
|
|
|
What is to be captured, the kinds of mixing and filtering, as well as expectations on interaction and visualisation all depend upon the target audiences initially identified by the 5 usage scenarios in section 1.
|
|
|
|
|
|
### 2.1.1 Participant Research Lab (S1)
|
|
|
Participant research involves capturing an activity addressing a specific research question with one or more people,
|
|
|
and then using both qualitative and quantitative analysis in an attempt to answer the research question or hypothesis. Examples of data commonly captured are: eye tracking, facial expressions, body language and verbal communication. Usually researchers sit in the room and take notes and survey questions are given to participants after the activity. Video footage is often reviewed but this is time consuming. The following forms of automated analysis of a scene may be of use:
|
|
|
|
|
|
* Gaze direction and duration of each participant
|
|
|
* Timing, duration and co-occurrence of facial expressions among participants
|
|
|
* The change in posture and activity of participants in response to events
|
|
|
|
|
|
In all cases the exact timing and sequence of, for example, gaze, is crucial for the researchers. Identifying patterns over time and correlations between events across modalities is challenging, doing this with multiple individuals interacting together is exceptionally difficult. In some cases a specific stimulus or event is introduced by a researcher and millisecond reaction timings of participants is important to capture. Therefore the needs of a participant research lab are:
|
|
|
|
|
|
1. sufficient quality to identify the measures of interest (gaze, facial expression, etc)
|
|
|
2. accurate timing and sequencing of events, with analysis synchronised to occurrence
|
|
|
3. consideration of privacy of participants when publishing data
|
|
|
4. record activity sessions that could last for a few hours at a time
|
|
|
5. multiple participants in the same activity must all be captured
|
|
|
6. both capture and analysis must be reliable and with minimal gaps in the data
|
|
|
|
|
|
### 2.1.2 Publicity Showcase (S2)
|
|
|
The University, department and research team require that this platform can showcase our ideas, expertise and competence to companies, funding bodies and other academic institutions.
|
|
|
|
|
|
## 2.2 DEPENDENCIES AND ASSUMPTIONS
|
|
|
We will rely, at least initially, on OpenCV and the Point Cloud Library. It is also assumed that we will not be using active depth sensors, although this should be a backup plan.
|
|
|
|
... | ... | |