Welcome to week 5, currently being posted in week 10. Funny how calendars work isn't it?
This week saw the tragic demise of a close friend. Microsoft's Azure Kinect DK. It was put to rest after a much-needed upgrade in the hardware department allowed for better running of something that felt impossible to understand.
- The system's guides were too high level for me to understand in such a short time frame, having had no prior experience in the field.
- The system requires hardware that is far too expensive for it to be viable to continue as a method of home rehabilitation motion capture.
The system has its uses, just not for our research. It is a fantastic device and I am by no means putting it down. However for the stage of research we are in, it is not as useful as eagerly anticipated.
I continued to work on my Python code for the remainder of the week, having gotten to a stage where data is accepted and something does indeed get spat out at the other end. The next part of the plan was to use a statistical measure to analyze collected data against the perfect motion path detailed in week four.
I opted for root mean square error, as I wanted an error based on the difference across the entire dataset.
This worked well, as we shall discuss in the sixth and final instalment of the research blog series.
Arron J Thompson