Brian Thorne - Computer Vision

Description

Computer Vision

Presented by Brian Thorne

Abstract

An introduction to computer vision in Python, from the general concept to its implementation with some current open-source libraries (opencv, scipy, pygame).

Outline

The purpose of my presentation is an introduction to computer vision (CV) and image processing in Python. I will aim to show what is possible with current libraries, demonstrate their usability, explore their limitations, and cover elements of best coding practice. Firstly, I will briefly present the basics, such as video grabbing, image processing and computer vision algorithms. I will illustrate my presentation with the Python bindings to the opencv library and show different examples such as displaying a video image, some filtering algorithms (dilation, gaussian blur), and how to implement facial recognition. I will describe after how related tasks can be achieved with other Python tools such as scipy and pygame. I will show how these three libraries can be used in conjunction to perform video capturing, image processing and user interfacing. Finally, I will introduce my basic Python computer vision framework, which is aimed at inexperienced users, and demonstrate it on the XO laptop, the inexpensive low-powered sub-notebook distributed as part of the One Laptop Per Child (OLPC) initiative.

[VIDEO HAS ISSUES: Sound and video are poor. Slides are hard to read.]