March 31, 2009 was our first night of combining the OpenCV Study Group with a weekly Maker’s Night Out session in the CCCKC Underground Lab. We changed things up for the OpenCV session with a presentation by Amith about fundamentals of image processing. He uses these techniques every day for medical imaging research.
Amith gave a brief comparison of medical imaging in a 3D environment using voxels where the 2D environment uses pixels to represent images. I’ve started reading the book Non-Photorealistic Rendering where a 3D model is used to represent surface textures of drawing materials in pencil sketches, pen and ink, and painting methods. Voxels are used there to simulate voids in the drawing surface where the art medium is deposited.
Amith spent a few minutes talking about video compression techniques and how a frame is displayed showing the entire image while subsequent frames only show changes from the previous frame, these are called P-Frame and I-Frames. The I-Frame or Intraframe is also called a keyframe and contains the entire image. The P-Frame is a predictive frame that only contains the changed data. A couple of nights earlier I came across a great example about How to Datamosh that plays with combining two video segments when most of the I-Frames are removed.
He also gave an overview of filters used in image processing and touched on a few filtering methods. For our next OpenCV session he would like to give some demonstrations of what various filters provide. As a challenge we are looking into methods of being able to use an image to identify coins and denomonations of coins.
Jestin was showing some of his recent work in OpenCV using color correction and filters to display live video. I had the prototype of a project I’m calling a Refrigerator Art Machine designed to capture artwork using OpenCV.
Others were busy tonight working on various Maker projects and crafts.