ConvnetClassifier

A complete guide to this application can be found here

Example showing how to train a model and do real time classification with ofxGrt and ConvNet.

This application requires the addons ofxCcv and ofxGrt.

Inputs are ConvNet activations as a 4096-bit input vector from the webcam feed.

Supports a long range of classification algoritms and allows users to save and load models.

Make sure the file image-net-2012.sqlite3 is located inside the folder ml4a-ofx/data, or change the line ccv.setup("../../../../data/image-net-2012.sqlite3"); to reflect the location of this file. The image-net-2012.sqlite3 file can be found in ofxCcv or downloaded directly from here.

To do:

• Run the convnet part in a separate thread

Instructions:

1. Click [record] or press [r] to record some training examples
2. Set the slider or press [1,2,3...] to switch class
3. Click [record] or press [r] to record new samples
4. Click [train] or press [t] to train the model