Ceramics + Object Detection


I’ve nearly finished the ceramic shells for the robots. They didn’t quite come together as I had imagined, but I think they still look good.

I was hoping for something a bit more asymetric and organic, but I didn’t have enough time to work on them before the clay dried too much to do drastic changes. I don’t have time to make anymore, so I’m going to have to work with these.

The object detection is going slowly but surely. I have recently trained my first “cascade classifier”. This is an XML file that describes the patterns that distinguish a particular object. The classifier I made was a bit too picky, so I need ┬áto adjust the parameters, and probably use more sample images. Here’s a tutorial which explains it a bit more.

I have decided two robots interacting with computer vision is going to be too complex, and I want to give myself plenty of time to make this look really good. Instead, I’m going to have the blood pumping robot react to a little simulation of micro-organisms on a TV. I’ve done something very basic like this before. I want to make the simulation inspired by the Nematode C. Elegans. This was the first multi-cellular organism to have it’s gene sequenced, and is the subject of an open source project to fully simulate it’s neural network.

Here is a cool timelapse someone made of C. Elegans developing:

Leave a Reply

Your email address will not be published. Required fields are marked *