We have a new web service that allows you to detect saccades/microsaccades in eye movement data using deep neural networks. You do not need to program the network algorithm yourself. Simply upload data and receive results from the world’s state-of-the-art algorithm for saccade/microsaccade detection!
To use the web service, follow this link and upload your data to train the deep neural network algorithm (which we call U’n’Eye). Then, with the obtained training weights, you can simply upload test data for classifying eye movements in these data. You will then obtain outputs indicating when saccades/microsaccades occurred.
If you want to use a pre-existing set of network weights instead, you can just upload the data for classification and skip the training step (although training is very fast anyway).
Details about the algorithm and its state-of-the-art performance are available in this paper.
The actual source code for the algorithm is also available here.