Objective
Understand how to write image processing applications with Python. Once done, the attendees can write more advanced and diverse applications (Intelligent robots, writing cool mobile apps, etc)
Audience
Machine learning and computer vision enthusiasts
Description
The OpenCV library is a powerful one - it comes with hundreds of built in functions. We'll see how to quickly get started with using the library and see some basic techniques in image processing.
Detailed abstract
Computer vision is one of the many aspects of machine learning. The applications of it range across a very wide variety: industrial robots, street cameras, sports robots, self driven cars, etc.
This talk will introduce to the audience how these ideas can be implemented. This isn't a workshop - but a quick stepping through the process.
Traditionally people use languages like C/C++ for such tasks. However, Python has the advantage of quickly prototyping your idea. The low verbosity of the language helps write idiomatic computer vision code.
In the talk, I'll go over the entire development cycle for a computer vision project: getting things setup, writing vision code, investigating if it works as expected/testing and deployment.
Outline
OpenCV (5 minutes)
Introduction to OpenCV as a library and it's capabilities
Interesting projects (5 minutes)
Some open source projects that use OpenCV to spark interest. (Soccer robots, augmented reality, panaromas)
Python + Opencv (10 minutes)
Python bindings for OpenCV and a quick look at getting up and running with the library. This includes installation of OpenCV and getting it working with Python.
Writing a prototype (15 minutes)
With OpenCV setup, I go through how to write scripts that process images. We'll write a script that tries to solve a Su Do Ku puzzle (after you've taken a picture of it)
Writing thumbnailer in Python + OpenCV (5 minutes)
A personal project I did - it generates better thumbnails of videos (for nautilus).
Alternatives to OpenCV (5 minutes)
Other libraries that are easier to use
Questions (10 minutes)
A Q&A session with the audience
Duration: 1 hour
Speaker bio
I've been working with image processing for over the past five years - mostly as a hobbyist and sometimes as a consultant for organizations and startups. I have a popular blog on image processing with OpenCV: AI Shack.
I'm a graduate from BITS-Pilani - Goa Campus. I'm currently working with DreamWorks Animation - where Python is the daily driver.
Links
- My blog on computer vision: http://aishack.in
- Personal blog: http://utkarshsinha.com
- https://www.linkedin.com/in/sinhautkarsh
1
▼
I'm running some maintenance on the http://aishack.in server - please use the wayback machine for now - http://web.archive.org/web/20140625133013/http://www.aishack.in/