lmccart / dynadash Goto Github PK

For my test in the lab, I have all 4 cameras pointing at me, then I press the "Analysis" or "Practice" push button, seems like at that moment the app determines the number of participants of the conversation via facial recognition, and ignore the "empty seats". But even if all 4 cameras should be able to see me (all pointing at me), sometimes maybe because of the angle or lighting or my head movement, I do not have 4 participants every time.... I do not know where the two physical buttons will be located on the table, but it will be even more difficult for the app to see my face if I am leaning over to press those buttons.....

What if after the button is pressed, the app will take a few seconds to "latch" faces. Then later on the faces may have movement and come and gone, but once "latched" any moment during the face-latching pre-session, it is considered a participant.

Or do you think it even make more sense to "latch" faces non-stop during a conversation? Any newly latched participate will just take zero initial status (zero interruption, zero dominance, etc)

The app will serial out the number of interruptions to the Arduino, as well as print it to the printer. But since it is just a single byte, that is 255 max. What happen if the interruption is more than 255 times? For the printer it is easy to still print the exact number. For the serial communication to the Arduino, what should we do?

At minimum:

• You interrupted people X times, more than anyone else in the group.
• You interrupted the most (xx times); Person A interrupted the least (xx times).
• You were interrupted X times, an average amount for this group.
• Person B was interrupted the most (xx times); Person C was interrupted the least (xx times).
• You spoke xx% of the time, less than anyone else in this group.
• Person B spoke the most (xx% of the time); You spoke the least (xx% of the time).
• You were smiling xx% of the time, an average amount for this group.
• Person A smiled the most (xx% of the time); Person B smiled the least (xx% of the time).

Icing. This data may also be desired -- we'll know more after testing with the "at minimum" above.

• You were smiling the most while Person C was talking.
• You were smiling the least while Person A was talking.