During the computational design course, last week, we explored a Processing library called OpenCV. The library provides tools (among the others) for object and blob detection through realtime video analysis. Today I was trying to figure out what are the core concepts linked to this technology.
OpenCV can detect faces (very well), eyes (ok), mouths (more or less), bodies (didn’t try yet) and the A-sign of the gesture alphabet (that it a closed hand-comunist-like-sign); that it means that the software can analyze the video and distinguish the mentioned “objects” among the rest of the video content. Every frame is geometrically divided into different areas (meaningful and not).
The software analyzes a video, so it detect objects over time. It detect movements. The time variable come into the play. Then, the software work in realtime. The way this technology is used depends on his characteristics, that I’m trying to describe here.
I found out different types of usage:
1. For visualizing graphics that adapt to the video content over time.
2. For using objects (in a wide sense) in a video and their movements as controller for a second piece of software.
3. For retrieving general data, such as how many people passed in front of the camera, that are collected through video analysis.
4. …
1. + 2. = For interactive games.
I feel those thought are a kind of unfinished and incorrect, because somehow I’m not able to go deeper and find the right way to describe the technology and its potential. By the way, I hope I was able to refine ideas around the topic.
