The art style of the users skeleton is plainly a ‘cartoonish stick man’ which they criticized it doesn’t match with the entire visual.
The initial idea of the users visual should be look like this:
During user test I changed it into ‘stick man’ visual because I realized Kinect v1 can’t capture high resolution image, and leaving noise edges around the silhouette which look like this:
It does not look ‘elegant’ with the entire visual, so I thought ‘stick man’ would be a better option. However the current visual is not match with the overlook and the concept, so I’m trying different user skeleton art style.
Firstly, I was thinking of doing Kinect puppet which skinned the user’s skeleton and get a puppet like animation and visual. However it consumed a lot of time to study the technical and a lot of calculation to do for one Kinect puppet. Looking at the time frame, I should spend more time on user experience and refine the entire visuals, hence, it still an unknown answer if this kind of visual really suit the entire look and feel or the technical will be stable or not. Therefore I leave this idea to next time.
I also think of ‘blob detection’ technique to stylize the silhouette, however the problem is the same: difficult technical and limited of time.
Considering of my rusty programming skill and limited of time left, I only can do subtle visuals on the user skeleton which is enough to look interesting. I still remember that I don’t want too much things happening on screen which will distract the users from understanding the core message of it. The first thing I think about is ‘Particles’ which it is easy to learned and flexible to modify. So here are the visuals effect I tried on.
*PS: the skeleton and silhouette visuals shouldn’t be covering up the screen during the actual situation.
I still want to stick with the black silhouette visual because it suit the art style and topic (which I want to present the idea of ‘loves has no colours and style to define’), therefore I try many types of particles style to match with the silhouette and at the same time ‘distract’ the users from seeing the low resolution of the silhouette. The particles makes the interactivity fun and looks abstract. Howbeit, there’s technical errors where I can’t make more than one particles on hands. Therefore I shift the ‘particles on hand’ idea to ‘interactive clouds or fog’, which might locate at the bottom or top of the screen, where users can move their arms to change ‘clouds’ direction and at the same time cover a bit of the ‘imperfection’ black silhouette. I will use this tutorial for this idea.