Gauri’s Status Report for 05/02

Last status report!  This week we wrapped up all loose ends of our project and finished integrating it all!  🙂  We also worked on our final presentation slides, I prepared for the presentation since it was my turn this time.  We think it went well.

We tried to retrain the classifier with gesture+head images to see if it would help solve the problem where we need a blank screen to make gestures in front of for it to recognize them.  It did not help and made the accuracy worse since there probably weren’t enough images like this in the training data.  We tested again however in slightly dimmer light conditions and switched the manual mode code to take 4 predictions into account rather than 2 (using all 4 processors) and it is actually working without the blank screen.  Yay!

Shrutika is now recording the demo parts for our final video and we will be finishing that up tomorrow.  Then, we will write our final paper and submit (shouldn’t be too bad since our design review paper was quite detailed).

Its been an interesting final semester 🙂