@PeRCeiVeLab   +39 095 738 7902


"Learning never exhausts the mind"

Deep learning human mind for automated visual classification

In our paper " Deep Learning Human Mind for Automated Visual Classification" (link) we describe a method for image classification based on EEG data.
We hereby provide:

  • The source code (using the LuaTorch Framework).
  • The dataset that we used for the experiments (EEG data).

Visit the page.

Gamifying video object segmentation

In our paper "Gamifying Video Object Segmentation" (arXiv link) we describe a web game architecture for crowdsourced collection of user clicks, to be employed for subsequent video object segmentation.
We hereby provide:

  • Access to the original game setup we employed for our work.
  • A public administration platform for setting up your own game and levels; please fill in the registration form to submit a request.

Knowledge Based Image Classification

We developed a tool for fast generating semantic enriched image annotations, which is suitable for every application domain that can be represented by an ontology. The tool has been extensively used to annotate our Fruit Dataset (3,872 images of 24 fruit varieties belonging to 3 different species), resulting in a large knowledge base (over 1,000,000 OWL triples about fruits, their context objects and attributes).
Visit the page.

"Nature is the source of all true knowledge. She has her own logic, her own laws,
she has no effect without cause nor invention without necessity." L. Da Vinci."