Engineering Thesis 2012-13
April 3, 2012 by admin

General Meeting 1

Spending time trying to get a similar C++ kinect base project working (I ran into problems/limitations with third party libraries with C#).

Having heard about how others working with the kinect have opted to base in C++, this is something that I will continue to work on.

In addition to being able to getting that set up, recording both audio + video etc. in a format that works with the third party libraries I want to use, is important so I can start drawing material from lectures to test.

  •   •   •   •   •
March 18, 2012 by admin

Presentation Skills

What the project will look for in presenters (aka. presentation skills), and how this would be feasible with third party tools + kinect sdk.

I read several books about presentation tasks. Notable books were:

  • Presentation skills for students, Emden and Becker
  • Powerful presenting, Vickers and Bavister

In summary, there are 3 areas to focus on: sound, movement, and direct non-verbal communication.

In regards to sound:

  • volume
  • pitch and variation (less monotony = better)
  • articulation (e.g. check for mispronounced t’s and h’s)
  • out of breath (gasping, trailing off at the end of sentences)
  • pace (observing audience feedback)
  • silence (pause before a number (not blurring words together), between sections, before asking if there are any questions and after having asked, avoid using filler words like UM and ERR and OKAY and SO )
  • trailing off at the end of the presentation
  • repetition of key messages with expression
  • using short sentences
  • apologizing
  • overuse of tentative words (maybe, possibly, try, suggest)
  • spending too long answering a question (waffling) and losing audience engagement

About movement:

  • Hiding from audience (behind a podium or constantly in a hard to see or shadowed location)/back to audience
  • Entrance into the room (walking speed, pre-speaking actions –  acknowledgement of audience coming in, cleaning boards in front of audience, avoiding signs of nervousness)
  • Signs of nervousness (looking down, fiddling, shuffling feet, tense)
  • Exiting (expressing exhaustion – sighing, sitting down immediately, disengaging from the audience too suddenly)
  • Introducing and engaging with next speaker if there is one (eye contact, smile with the next speaker)
  • making use of the primary effect (beginning and end of presentation where audience is most engaged) – full engagement, most clarity and emotion – e.g. what could be/something shocking, the aims and agenda, conclusion
  • standing position (feet good distance apart, not slumping, good posture, and not in front of the other or inwards etc.)
  • Pacing the stage (not while talking, use to change the room energy between points in a pause)
  • Resting hands position (avoid: behind back, too high, in pocket, over groin or hips, folded,

And direct non-verbal communication:

  • Smiling
  • Engaging with audience (eye contact for good lengths of time, looking at different areas of the room, responding to visual cues from the audience (e.g. disinterest, confused), repeating and speaking to whole audience when answering questions)
  • Confirming the question was answered with the original question asker
  • Re-engaging when audience attention lessens (e.g. asking question, and answering before audience can answer)
  • Occasionally gesturing for emphasis (e.g. hands, moving forward) to support something emotive/important, or asking question.
  • Using open hands (e.g. to indicate someone in the audience), not closing palms on anything
  • Effective gestures (not too close to body, above elbow height, don’t show palms upwards)
  •   •   •   •   •
March 7, 2012 by admin

First goals

During this meeting, we created some initial goals for the project:

  • Detecting if the lecturer has his back to the audience
  • Use of filler words when an issue occurs (um… err…)

Also, introduced to openear, opensource project to read emotion, which would have the trained models already made. However, as a C++ library, I need work out how to develop in C++ with the kinect, or find some way to use the C++ functions from C#.

Another goal is to be able to record a lecture and encode it into a smaller format such as H.264, as there are difficulties with frame + audio syncing.

  •   •   •   •   •
February 26, 2012 by admin

Lecturer request

Received a suggestion from a lecturer that they wanted to see audience reaction, and display in a similar way as during the pm elections (with a worm graph/polygraph)

http://news.bbc.co.uk/2/hi/uk_news/politics/election_2010/8636745.stm

The pm election system used an electronic pad for voting, but the kinect could be used to determine audience emotion, this could be automated into a graph for lecturer feedback.

It would be a good addition to something similar to the sparkline concept from the book “resonate”, Duarte, for feedback after the presentation along with video playback.

 

  •   •   •   •   •
February 19, 2012 by admin

Tools for presentations

What tools would be useful for a lecturer or presenter? After doing some research, the most useful tools would be related to equipment automation/control (to save the lecturer time), as well as audience feedback for the lecturer in real time.

For example, in this article:

University of California – San Diego. “Computer Scientist Turns His Face Into A Remote Control That Speeds And Slows Video Playback.” ScienceDaily, 25 Jun. 2008. Web. 29 Jan. 2012.

Facial expressions are used to control the speed of video playback, the analogy being that if the viewer is confused, they need more time to think about it, or the viewer could be bored and the pace is too slow. Ways to determine this could include blinking, nodding and facial expression.

Whitehill, J., Bartlett, M., Movellan, J.: Measuring the perceived diculty of a
lecture using automatic facial expression recognition. Technical Report 2008.01,
Machine Perception Lab (2008) – http://mplab.ucsd.edu/~jake/its08.pdf.

Uses 13 different “Facial Actions” to analyze a students’ engagement in a lecture. A strongly related expression to difficulty was found to be blinking (less blinking in difficult segments).

“Inauguration Of The World’s Most Sophisticated Electronic Classroom.” ScienceDaily, 23 Nov. 1999. Web. 29 Jan. 2012.

The need to outsource the low-level device control so that the lecturer can focus on giving an effective lecture. The kinect could provide some room management by actions.

  •   •   •   •   •
February 15, 2012 by admin

hello kinect

Showing an IR view and a normalized view of the depth map, tilt and color/filter controls.

I have decided to switch to using the Kinect SDK and starting over because:

  • There is a detailed skeleton feed (body tracking)
  • Access to the audio array on the kinect (not currently possible with opensource solutions)
  • Frame events and better support/communication with the kinect

Starting the learning process for VS, C# and the kinect sdk using the kinect sdk quickstart tutorials.

  •   •   •   •   •