Here’s a video of me playing with my Gestural Music Sequencer. I’ll upload a better version at some point, but I think you can a least get a sense of how you might use this kind of tool (I’m talking about the sequencer, not the performer). As you can see the video has been mirrored so it’s easier to follow your own movements.
To reiterate how the sequencer works, the X axis of the brightest pixel is used to determine the pitch, while the Y axis is the dynamics. The application outputs MIDI data that I’m routing to Reason. I’ve programmed the up and down arrows on the keyboard to increase or decrease a multiplier that along with the 15 fps frame rate determines the time between each note on event. I’ve also enabled a group of keys to adjust the transposition.
I have recently added keyboard controls to dynamically set the transposition in my Gestural Music Sequencer. I set it up like a piano keyboard, so that “Q” is the key of C, “2” is C sharp and so on. In this example I have the application set to to play a minor pentatonic scale. This makes it quite clear when the key changes. In this case I’m modulating between C and F sharp. I also started with a slower note duration, then switched to a faster one. I would not classify this as a very musical performance. This was part of a test to make sure that all of the transposition key controls were operating properly. 
As I mentioned before I have GMS setup to produce specific scales. At this stage they are all based on the key of C. Eventually I’ll set up the application so that the key and scale are dynamically adjustable. I will also include a wider variety of scales, including all the modes, diminished, whole tone, and more. 
The intervals in a major scale are really the same as a minor scale. What’s different is the starting point. All the pitches in a major scale are the same as its relative minor, which is down a minor third. So the difference between a melody in a minor key and a major one is the root note. The same melody can be minor or major depending on the perspective of the listener. The last pattern I posted is only in a minor key if you consider C the root. If you consider E flat the root, then the melody is in a major key with C being its relative minor. In this case, the pattern is in C major, so if you listen with an A (the relative minor of C) in mind you’ll hear it in a minor scale. 
In my previous example of audio created with my gestural music sequencer, that I’m tentatively naming GMS, I presented a pattern of sampled Rhodes notes in a chromatic scale. One of the functions I’ve built into the application is the capability of switching scales. Currently the available scales are major, minor, pentatonic minor, and chromatic. Here’s an example of the application producing notes in a minor scale. One thing you may notice is the dynamic range. By gesturing lower within the Y axis the notes get quieter, while gesturing near the top of the screen makes notes that are louder.