| Hey everyone! I continue to play with clouds and today it’s time to reveal the CloudCV - a cloud-based image processing project. | Based on my previous posts i host a server in the Digital Ocean’s cloud. | I have to say, everything is working like a charm. | The cheapest 5$/month plan gives me whatever i may need for this project. | All the source-code is already sits on Github and you are more than welcome to study it.
Vacation time is over, and now i’m on my way from Tartu, Estonia where i participated in 48 km. inline speedskating marathon to Odessa. My bus have Wi-Fi onboard, so i decided to write a short success-story how i managed to build a C++ addon module for Node.js and run it on the real server inside the Cloud9 IDE. You may also want to check the first tutorial since this guid relies on it.
A long time ago i was playing with cloud-based image processing. The first reason why i didn’t shared a reciple how to compile OpenCV as native app for windows azure cloud was trycky build process. It was too complicated and this tutorial will become outdated very quickly. The second one - Azure hosting wants a lot of money. So i put my research in this area on hold for better times.
img.pull-left.img-thumbnail(src=“arbasketball-logo.jpg”,alt=“ARBasketball”) p ARBasketball was one of the first augmented reality-based games in App Store. It has been published in 2010. In these days not many people have even heard about AR. I mean it wasn’t so popular as it became now. But there were people who saw the great potential in this growing market. One of them was Konstantin Tarovik, the author of ARBasketball. I must confess - I saw this application before, but had no idea it’s author lives in Ukraine, and in the same city as I am!
This post is outdated. Please, visit updated post: Integration of KAZE 1.6 in OpenCV A new version of KAZE features has been integrated my private fork of OpenCV (You can find it’s here: https://github.com/BloodAxe/opencv/tree/kaze-features). We’re on the way to make pull-request and integrate KAZE features to official OpenCV repository. There only few things are left: Include KAZE into features2d unit tests. Rewrite KAZE to support OpenCV threading API.
OpenCV library is widely used by computer vision engineers across the world. It contains almost all algorithms you may want for R&D or product development. It has production-ready build farm with tests and strong community that give nice feedbacks and discover errors. But nevertheless OpenCV has some strange issues and undocummented behaviour that can surprise you as minimum and crash your app as maximum. How to get diagonal matrix in OpenCV A typical parameter update computation in non-linear optimization using Levenber-Marquardt algorithm looks like this:
Recently i came across the publications to a new features called KAZE (Japanesee work meaning “Wind”). They interested me, because KAZE authors provided very promising evalutaion results and i decided to evaluate them too using my OpenCV features comparison tool. Fortunately KAZE algorithm is based on OpenCV, so it was not too hard to wrap KAZE features implementatino to cv::Feature2D API. This post is outdated. Please, visit updated post: Integration of KAZE 1.
During my research and development work i often have to display a lot of text infromation on top of the OpenCV images. You know what i mean. Suppose you’re writing video stabilization algorithm. On each frame you want to display number of features visible on current frame, number of features matched with previois frame, camera motion parameters, recent twitters, name of your pet, etc.. In the OpenCV you can use cv::putText function to print formatted std::string at the desired position on the image.
I feel excited writing this post. The “Mastering OpenCV with Practical Computer Vision Projects” books is done and published! This is my first experience as a book author and i hope you will like it. This book has been written by several authors, and covers many topics that has real appliance in computer vision area: augmented reality, face recognition and head pose estimation, structure from motion estimation and cartoon image processing.
I continue playing with powerful NEON engine in iPhone and iPad devices. Recently i bought iPhone 4S that replaced my HTC Mozart and i decided to check how to speed up BGRA to GRAY color conversion procedure using multithreading. Recently Itseez announced a minor release of OpenCV 2.4.3 with a lot of new major features: Added universal parallel_for implementation using various backends: TBB, OpenMP, GCD, Concurrency Improved OpenCV Manager, new Java samples framework, better camera support on Android, opencv2.