It is a neural network that tries hard to determine what object is shown in the picture. Due to that it is just a simple single-layer perceptron it can only work with two classes of objects, but I thought it might be nice to have something like this, that works with images.
Also it is my first neural network, so please, feel free to criticize, I need it!
First you'll need to compile the whole thing. The main class is ImageRecognition.
javac ImageRecognition.java
To train this perceptron you need to have a set of images for each one of the two objects.
So to start the training you need to create two folders Object1 and Object2 and move corresponding images to each folder. Then you need to do the following:
java ImageRecognition -learn -Object1Name Object1/* -Object2Name Object2/*
Next, you can use this trained perceptron to recognize these objects in other images:
java ImageRecognition [path_to_image.png]
Or reset the perceptron to retrain.
java ImageRecognition -reset