Clip Documentation

How to use Aria

Everything You Need to Know about Finger Gesture

Gesture Concept
When we talk about finger gesture, you might already have a lot of different gestures in mind. The gesture of thumb up, the gesture of rock, the gesture of shaka, the gesture of love, the gesture of victory and many others.
If you think of any of these gestures, you should be able to picture how these gestures look like: they look like symbols! They are static position of the hand where the fingers don’t move.

However, the finger gestures in our context has an opposite focus.
We want you to move your fingers!
So the gesture looks more as a motion, thus instead of asking what’s the shape of your gesture, we would ask “what’s the move?”

Some other simple finger gestures that might be seen in daily life could help you immediately understand this different focus.

Gestures like flick, snap, tap... when mentioning them, you probably see the motion of the gestures instead of one shape, isn’t it?

Option 1: we introduce now the 4 gestures
Still, we strongly suggest you to start with simple gestures like flick, snap and tap so that you understand the requirements for the gestures.
(If we are going to stick to this option, then the examples in the component part should be changed to two among these four)

Gesture Components
In the previous section, we said that there are certain requirements for the gestures.

In order to better explain those requirements, we need to show you how we breakdown the gesture movement.

We will use the index flick and the index tap as examples:

The whole movement of a finger gesture consists of three parts:
  • the head part
  • the core part
  • the tail part
We will firstly talk about the core because it is the most important part and it helps us to identify the remaining two.

  • Core
The core of a gesture is the most important part for our algorithm, because it’s the highly recognizable part. Sometimes we call it the gesture signature.

For an index flick, it is the part where your index shoot out from behind your thumb. For an index tap it is the part where your index rapidly knocks into your thumb.
  • Head
The head part of a gesture is the movement that happens before the core. It is normally the movement necessary to generate the force.

For an index flick, it is the part where you place your index behind your finger.
For an index tap, it is the part that you raise up your finger preparing for the knock.
  • Tail
Contrary to the head part, the tail part is the movement that comes after the core.
It is normally the part where you relax the muscles on your fingers and bring your whole hand back into a relaxed position.

For an index flick it is the part where you relax your stretched out index and bring it back to relax position.
For an index tap it is simply the part where you relax you thumb and index so that they are not pressing against each other.

Gesture Requirements (What & Why)
You might wonder why there are some requirements for our gestures..
Here we try to get you a first insight on our technology:

Theoretically when you repeatedly perform the same gesture, each biomechanical signal that you produce should be always the same.

In reality this is true only when you reach a level of highly repeatability of the gesture, so these requirements are meant to get you as close as possible to the optimal performance.

This means that following our requirements, you will be able to transform your gestures in a unique but coherent template that can be always recognised by FlickTek algorithm.

  • A strong core
The first requirement is that the core movement should be strong. Which means you should put enough force into the gesture.
The reason behind this is that we want a rapid movement of your finger and normally the stronger you perform it, the faster your fingers move.
  • Light and short head and tail
The second requirement concerns the head part and the tail part.
For both of these two parts, your movements should be light and relaxed.
Try to avoid any unnecessary movements!
  • Examples of bad performances
Even though most people naturally perform the gestures following our requirements, we do notice some typical errors.
You should check if you are accidentally doing the gesture like these.

- For a flick like gesture (index flick as example), in order to perform a gesture strong enough, some people would keep pushing the index against the thumb for a while before they shoot out their index.
You should try to avoid prolonging this pushing, try to make the process of pushing and shooting as smooth and continuous as possible.

Another typical error while doing a flick is that after shooting out the index, some people tend to keep it stretched out and the muscle tighten. This should also be avoid. You should relax your fingers immediately after the shooting.

- For a tap like gesture (index tap as example), as a result of performing a strong tap, some people would keep the index and thumb pressing against each other after the knocking happens.
You should try to avoid this extra force applied on your fingers.
Immediately after the knocking, you should relax your fingers, but please don’t over correct this as forcing your fingers apart, just relax and bring your fingers to the comfortable position slowly.

  • HINT:
You should try to replicate every details of your gesture.
While your fingers touch, try to always touch the same point.
While you apply force, try to apply the same force etc.

Record your gesture

Everyone’s finger gestures generate unique biomechanical signals. Therefore you will need to record them in the Clip so they can be recognised.
And to make sure the algorithm finds the repeatable pattern, you will be asked to repeat the gestures for several times until the algorithm is confident about what to look for.

The followings will be a description on how to use the App to record the gestures.
Since everyone does their gestures differently, we need to show Clip the gestures that you are going to perform and show it for enough times so when you do it again in the future, Clip will be able to recognise them. And this calibration process (record your gesture) is where it happens.
When you press start, you will see the screen telling you which command this gesture is going to be linked to. You need to choose a gesture that pass all the training and to perform for five times. After every repetition, you will see the screen update to provide you feedback. Normally, you should just be informed that you are getting closer to get this gesture done. But you may also receive two different reminders.
The first one tells you that this current recording is very far from you previous ones. You need to do them again.
The second one tells you that at least one of your previously finished gesture has data pattern too similar to the current one. You should try to avoid this new gesture or we might mix them up.

Testing App: check your skills!

When you finish all you four gestures, you are good to use Aria. But before that, you could also test if Aria is responding your gesture correctly by testing them out.
Inside the Testing App you can:
Freely check that your gesture are well recognized
Run a test where the given a random sequence of gestures, you’ll be asked to perform them one by one and then get a score based on your performances.