Making The Computer See: Computer Vision For Everyday Applications
With a camera and a powerful computer in every pocket, the shift from typing to computer vision is just about to happen.
Instead of entering information about a wine for our tasting notes, we snap a photo and let the computer recognize it and look up the information.
Instead of having the shopper enter his credit card info, we can read the information straight from the card with the camera.
We can even use computer vision to play a fanfare when someone arrives at our doorstep carrying a pizza box.
The tools to make this happen are in the field of computer vision. Today, the are solid techniques and cross-platform open-source libraries such as OpenCV available that make it easy to build this into everyday applications.
This presentation will show you how.
We will look at practical applications of computer vision such as using the camera to scan text (OCR), reading the info from a credit card or a license plate from a passing car, recognizing the pizza box in the image or using the wine label as a visual query against our wine database to find the producer, vintage, grapes and other information with no manual data entry required.
The talk combines practical examples with a presentation of the basic image processing techniques that make it possible.
You will learn about basic image transformations, how to find interesting or characteristic parts of image, extract and recognize text and and ways to compare images and how to recognize known shapes and objects in images.
Be prepared for a highly visual talk that provides not only an introduction to making the computer see but also presents its mathematical and statistical machine learning underpinnings in a very practical context.
Developers curious about how to max out the CPU and camera on their smartphone while saving their users a lot of trouble.Entry level talk: No Ph.D., machine learning or computer vision background required.
Carl Franklin has done extensive work with the Kinect for Windows SDK v2, and now he's sharing his knowledge and code with you. He shares code to simplify Kinect development as well as GesturePak Alpha v2, which lets you record and recognize gestures in C#.
Have you heard? Apparently we’ve created a dreadfully insecure internet with vulnerabilities reaching so far and so wide that literally anything is obtainable online through covert methods. Often this involves the now very well-known yet frequently present classic exploits – SQL injection, cross site scripting and others – but now we’re also seeing new attacks against security defences such as two factor authentication.
In this session I’ll take you through how the risks we, as developers, are building into web sites and APIs can be easily exploited to gain access to everything from credit cards to credentials to control of commercial facilities. For many people, they’ll be stunned at the simplicity of the risks that continue to be exploited whilst for others, risks they never knew existed will be exposed, decomposed and most importantly, the mitigation will be shown.
This session recreates real world examples of attacks against airlines, ticketing systems, hotels and transportation services – enough that someone literally could hack themselves all the way around the world to Norway. It’s not a theoretical exercise; these are real world attacks by real world hackers laid bare.
Since the introduction of lambda expressions in C#, we have had two different style of programming. Yet, programmers used to the habitual style often find it easy to fall back on those old practices. In this presentation we will take a number of common tasks we code in C#, discuss the downsides of the habitual style, transform it into functional style, and discuss the benefits. We will also discuss some techniques that can help make this transformation easier on everyday projects.