We have released a few patches for bugs reported for a release 2.3
The NetBeans project wth all patches is available for download here:
https://sourceforge.net/projects/neuroph/files/neuroph%202.3/neuroph_2.3.1_nb-RC1.zip/download
Patches include:
Neuroph v2.3 - Image Recognition has been released! The most exciting feature of this relase is image recognition support which provides GUI tool for training multi layer perceptrons for image recognition, and easy to use API to deploy these networks in the end-user applications. With this IR library you can perform image recognition in just few lines of code, simple as:
HashMap<String, Double> output = imageRecognition.recognizeImage(new File(”someImage.jpg”));
Step by step tutorial is available here
In addition to image recognition this release also comes with:
We have decided to re-schedule the release of NEAT support until we polish it (although we announced it), so it will be released officialy with the next release or as a separate component.
We have completed the image recognition support for Neuroph. We’ve created easy to use GUI tool and API, so developers can easily add image recognition with neural networks to their applications. The image recognition support will be released in public with the next release of Neuroph 2.3.
Thanks to Aidan Morgan, the Neuroph got the NEAT support!
Aidan joined the development team and made the great contribution, the support for Neuroevolution of augmenting topologies, the technique which use the genetic algorithm to evolve the neural network.
This contribution will be released in public with the next release of Neuroph 2.3.
Two-part interview titled Neuroph: Smart Java Apps with Neural Networks has been published on developers social network NetBeans Zone. Take a look at part 1 and part 2 and and get to know the Neuroph project behind the scene, as well as our future plans. Part two of the interview also contains very brief howto for the Neuroph GUI.
The latest version of Neuroph Framework has been included in Gentoo Linux package repository. This is a great honour for our project and it will help to grow the Neuroph community. Thanks to Alistair Bush for this. Read more about Gentoo Linux here
http://packages.gentoo.org/package/dev-java/neuroph
http://packages.gentoo.org/package/dev-java/easyneurons
Neuroph v2.2 has been released publicly! This release brings:
With all these new features like momentum, import training set from files and network error graph, Neuroph will be more suitable to use for many practical tasks and easier to experiment with.
The sample image recognition application within this release has been used to train the neural networks for DotaScript
This release also comes with full NetBeans project tree which will make easier to develop and build project according to its original structure.
We have almost prepared the new release 2.2 which we hope to make public by the end of the June, and it will bring:
An article about how Neuroph is used in DotA Script for image recognition has been published on Slashdot! read the article at http://games.slashdot.org/article.pl?sid=09/05/14/0447244
This was a big step for promoting our project since project site got over 3000 visits and 500 downloads in just 3 days. We’re also very happy to see that there are many people having positive thoats about our project.
The new version of DotA AutoScript has been released. This script, used as utility in game playing is based on the newest additions in Neuroph: image recognition library and momentum backpropagation learning.
Script is being used to recognize the on screen images in real time during game play, and this is a good example of how neural networks can be used for image recognition.
“There are two different types of neural networks used by DotA AutoScript. The first type is a simple binary image classifier. It uses Neuroph’s “Multi-Layer Perceptron” class to model a neural network with an input neurons layer, one hidden neurons layer, and an output neurons layer. Exposing an image to the input layer neurons causes the output layer neurons to produce the probability of a match for each of the images it has been trained to identify; one trained image per output neuron. The output is “binary” in the sense that an image should be considered a complete match, or not at all. DotA AutoScript uses a threshold of 80% probability to confirm a match.”
Copyright © 2009 Geeknet, Inc. All rights reserved. Terms of Use