Weekly Update 3/31/2011

What did we do this week:

This week more effort was put into merging the server and the hand tracking.  We were finally able to succeed on Monday night.  This allowed us to use the system to conduct a user study on Tuesday and have the report written for Thursday.  The study revealed a lot about our system and gave us some good ideas about where we need to go in the future.

Problems:

When running the user study we found that our clicking mechanism was pretty bad to say the least.  Users had no precision with their clicks at all because the hand would move while the clicking gesture was performed.  Multiple solutions have been proposed to fix this.

Project Status: 

The project is moving along nicely now that we have some of the proof of concept programs integrated.  We can now fix the clicking problem and look into gestures that can be added to our system.

New Ideas:

The clicking was a problem so we are looking at two solutions.  The first would be to write our own push recognizer that would take the click sooner than the NITE push recognizer.  The second is to back track through inputs to where the click started and take that point instead of the point where the click was recognized.

Weekly Update 3/24/2011

What did we do this week:

We have a lot of proof of concept programs laying around that work independently but only do one or two things. This week we focused on bringing some of these programs together.  The primary work was done integrating the Kinect hand tracking software into the background server.  These two programs, when properly merged will provide the backbone for the final product. 

We also received another Kinect yesterday which will allow us to streamline the development and user studies.

Problems:

The merging of the programs is moving a little slower than expected.  There were some roadblocks regarding dependencies and missing libraries that are slowly being fixed.  This isn't a show stopper though, and should be overcome shortly.

Project Status: 

The project is a little behind and will take some effort to get back on track if we want to implement the total design by the end of the semester.  Fortunately the work for the next couple weeks should be somewhat easier and therefore more enjoyable.

New Ideas:

No new ideas for this week.  The current implementation issues are taking precedence.

No important picture for the week as no new features have been implemented. All effort has been focused on consolidating all our proof of concept programs.

Weekly Update 03/10/2011

What did we do this week:

The gesture recognition proof of concept's progress has slowed due to external factors however we were able to get the system mouse to be controlled by the Kinect with clicking by pressing in on the hand.  This was a major hurdle for us.

Problems:

Some of the biggest problems we face right now are external issues that are occupying our time, keeping us away from the Kinect project.  Everyone in the group got hammered with homework from other classes and we had a slew of tests because it is the week before spring break.

Project Status: 

Despite the external distractions the project is coming along quite nicely.  We now have the basic framework down for controlling the system mouse and clicking.  The next step is to progress into recognizing gestures.

New Ideas:

No new ideas for this week.  The current implementation issues are taking precedence.

Weekly Update 3/2/2011

What did we do this week:

Aaron started working on an application to recognize gestures form the Kinect.  Andy was able to complete the framework for the background service, including the data passing part.  Warren looked into getting hand tracking data points from the Kinect and was able to get raw data.

Problems:

We are at a hump that we need to get over.  There are many pieces of the puzzle that now need to come together.  This will involve some oversight and planning.

Project Status: 

We are on a good pace right now.  I would say we are a little paused due to other schoolwork but this should not be an issue much longer.  This weekend we hope to get a substantial amount of work done.  We would like to be able to control the mouse using the Kinect.

New Ideas:

No new ideas for this week.  The current implementation issues are taking precedence.

Items for Testing

From Service:

• Open socket successful
• close socket successful
• bind on port
• test connection sequence
• test connection success
• test connection refuse
• messages formed correctly
• messsages sent across the network correctly
• client recieves messages
• client process messages(test callbacks)
• client connection timeout
• client shutdown connection
• server shutdown connection
• server unexpected shutdown handled
• correctly chooses client that has focus
• stress test - many clients
• stress test - many messages

From the Kinect:

• Automated tests can be established that uses a recording of someone at the Kinect.

Weekly Update 2/24/2011

What did we do this week:

This week we got a hold of some Kinects so we were working on getting the sample projects to compile.  We read documentation regarding OpenNI and NITE to help understand the modules a little better.

Problems:

We are now facing the problem where we don't really understand the system and the modules enough to work on it full speed ahead.  

Project Status: 

The project is finally starting to pick up some real steam.  For the first week we have actually accomplished something regarding the Kinects.  We are now able to, for the most part, run the samples meaning almost everyone has the drivers installed and working on their computers.

New Ideas:

We are trying to come up with a good data model for passing gestures between the service and the third party applications.  We probably need to know more about what data we will be getting from NITE before we can fully define this though.

Weekly Update 2/15/2010

What did we do this week:

This week we started working to install the Kinect drivers.  We are still waiting on a Kinect so progress here is minimal.  The team worked to try and understand the OpenNI/NITE framework that we will be using.  Documentation is not very good on the frameworks so a lot of energy must be spend attempting to learn the frameworks.

Problems:

Just like last week, our biggest problem is that we don't have an actual Kinect.  We finally got a Kinect on Wednesday so we can finally start working on the project.

Project Status: 

The project is finally picking up.  It had been stalled all night but now that we have one Kinect we can start slowly working on the project.  More Kinects will allow faster development.

New Ideas:

Due to the relatively low amount of work being done while we wait for hardware there have been no new ideas about the project.

Next Week:

Planned:

We hope to have hardware this week so we can begin to start getting our hands dirty working with the actual Kinect. 

Goals:

If we can get hardware then everyone should have the drivers successfully installed.

Video: