ViCoS Eye is an experimental online service that aims to demonstrate a state-of-the-art computer vision object detection and categorization algorithm developed in the Visual Cognitive Systems Laboratory at the Faculty of Computer Science and Informatics, University of Ljubljana.
The purpose of the ViCoS Eye is twofold:
- to bring the current state of computer vision research closer to the people (both the capabilities and the limitations),
- to give us, the researchers behind the algorithm a better insight into the main issues that cannot be noticed by simply using the standard benchmark datasets used in standard performance evaluation.
What can it do?
At the moment you can upload an image and the system will detect known objects in it and detect and/or categorize them. With time we plan to add more functionality like searching for similar images.
The algorithm has been trained on selected images from a standard benchmark datasets Caltech 101 and ETHZ Shape Classes, therefore its knowledge is limited. With time we plan to extend this initial knowledge database by also incorporating submitted images to improve the overall performance of the service.
The detection and categorization is at the moment limited to two categories: mug and Apple logo.
The entire image categorization supports more categories:
- stop sign
- dollar bill
- scissors (upside oriented only)
- soccer ball
- ceiling fan
A short history of ViCoS Eye
- July 2012 - The idea
- August 2012 - First prototype created
- September 2012 - Android app protorype
- December 2012 - Service rewritten as a Tornado server
- January 2013 - Alpha test (categorization of entire image)