Automatic image segmentation techniques based on supervised machine learning algorithms

The goal of this project is to create a technique which can automatically segment images based on extracted image features. With the help of supervised machine learning algorithms, (in this case the Random Forest Classifier) there is no need for human interaction. While several image segmentation techniques exist; for example magic wand tool in Adobe Photoshop, they usually require manual interaction.