Results **new** 2003/12/16 14:54:07.203 US/Central 2003/12/16 14:54:24.730 US/Central Sivakiran Nagisetty nsiva@rice.edu Sivakiran Nagisetty nsiva@rice.edu Results from the testing Using the Matlab command ginput to isolate the mouth from an image and then performing tests to detect mood we had the following results.Results with user defined cropping Subject # Input Sequence Output Accuracy Subject 1 Angry, Happy, Sad, Surprised Angry, Happy, Sad, Surprised 100% Subject 2 Happy, Surprised, Sad, Angry Happy, Surprised, Sad, Angry 100% Subject 3 Surprised, Sad, Angry, Happy Surprised, Sad, Angry, Happy 100% Subject 4 Sad, Happy, Surprised, Angry Angry, Happy, Surprised, Sad 50% Subject 5 Angry, Happy, Sad, Surprised Angry, Happy, Sad, Surprised 100% Subject 6 Happy, Surprised, Sad, Angry Happy, Surprised, Sad, Angry 100% Subject 7 Surprised, Sad, Angry, Happy Surprised, Sad, Angry, Happy 100% Subject 8 Sad, Happy, Surprised, Angry Sad, Surprised, Happy, Angry 50% Subject 9 Angry, Happy, Sad,Surprised Angry, Happy, Sad,Surprised 100% Subject 10 Happy, Surprised,Sad, Angry Happy, Surprised,Sad, Angry 100% Subject 11 Surprised, Sad, Happy, Angry Surprised, Angry, Happy, Sad 50% Subject 12 Sad, Happy, Surprised, Angry Angry, Happy, Surprised, Sad 50% We then ran the test using the function goodcrop instead of doing the cropping manually using ginput. We obtained the following resultsResults using goodcrop Subject # Input Sequence Output Accuracy Subject 1 Angry, Happy, Sad, Surprised Angry, Happy, Sad, Surprised 100% Subject 2 Angry, Happy, Sad, Surprised Angry, Happy, Sad, Surprised 100% Subject 3 Angry, Happy, Sad, Surprised Sad, Happy, Angry, Surprised 50% Subject 4 Angry, Happy, Sad, Surprised Surprised, Happy, Angry, Sad 25% Subject 5 Angry, Happy, Sad, Surprised Angry, Happy, Sad, Surprised 100% Subject 6 Angry, Happy, Sad, Surprised Sad, Surprised, Happy, Angry 25% Subject 7 Angry, Happy, Sad, Surprised Angry, Happy, Sad, Surprised 100% Subject 8 Angry, Happy, Sad, Surprised Angry, Surprised, Happy, Sad 25% Subject 9 Angry, Happy, Sad, Surprised Angry, Happy, Sad, Surprised 100% Subject 10 Angry, Happy, Sad, Surprised Angry, Happy, Sad, Surprised 100% Subject 11 Angry, Happy, Sad, Surprised Angry, Happy, Sad, Surprised 100% Subject 12 Angry, Happy, Sad, Surprised Sad, Happy, Surprised, Angry 25% The overall accuracy of the mood detection algorithm , when using the matlab function ginput, was 83%. The overall accuracy when using the goodcrop routine was 71%.