🎉 🎉. RAPIDMINER 9.8 IS OUT!!! 🎉 🎉

RapidMiner 9.8 continues to innovate in data science collaboration, connectivity and governance


How do i get image classification model designed and working with RapidMiner 9.7.002?

SridharBSridharB Member Posts: 1 Newbie
edited September 15 in Help
I have the following extension installed 
–Python scripting extensions
–Image Handling extensions 
–Deep Learning extensions
–Keras Extensions
–ND4J backend extensions

I am trying to run a NIH image classification data set with RMStudio. The layers in deep learning was reused as highlighted in the deep learning tensor example.  I am unable to have Turbo Prep and AutoML to work with this data set. Am i missing something? 
Any pointers to design new models and publish as an operator will help( this for within the IDE not marketplace)


  • jacobcybulskijacobcybulski Member, University Professor Posts: 269   Unicorn
    If you already have Image Handling and Deep Learning extensions, it should be quite simple to read in your directory of images with IH. Unfortunately, with DL extension you cannot supply a separate x and y vector for the DL classifier. However, when you organize your images one class per image folder and use a class name as a folder name, you can read all images in with the folder name as a class name (use a label option in Read Image Meta Data operator). Once all images are read in and labelled you will be able to create a convolutional NN classifier without Python. To make things easier, find the Add Convolutional Layer operator, look at its Help, and open the associated tutorial on MNIST. Note however that MNIST files are not provided but if you structure your folders with one sub-folder per image class, you will be able to classify your images without many issues.
Sign In or Register to comment.