Dipartimento di Informatica

Date: Wed 11 Nov 2020
Time: 9:00
Duration: about 20 minutes
Venue: meet.google.com/xqj-mswe-wbv
Speaker: Danilo Avola

Title: Computer Vision and Non-Visual Information Supported by Artificial Intelligence for Smart Systems and Emulation of Human Skills

Abstract: In the last two decades, Computer Vision and Artificial Intelligence have made great strides in an ever-increasing number of complex tasks in research, civil, and military applications. Some examples are Earth Observation at local and global scale, where data is acquired by sensors on geo-stationary satellites; Scene Understanding, where data is collected via static or Pan-Tilt-Zoom (PTZ) ground cameras; and Land or Underwater Monitoring, where data is obtained from sensors on Unmanned Aerial Vehicles (UAVs) or Autonomous Underwater Vehicles (AUVs). Common ground of the aforementioned systems is to be able to observe, understand, and act (e.g., feedback, action, reaction, alarm, and so on) without any human intervention. Specifically, the objective is to obtain autonomous and smart systems capable of dealing with complex events. To reach this end, it is sometimes necessary to take a step further, moving from smart systems to systems able to emulate human skills. Instances of such applications are, among many others, drones able to decide where and how to look for a target, humanoids able to perceive lies and emotions by humans, and systems able to re-identify people and their actions. What is more, visual information, in some contexts, can be enriched, or even replaced, by different non-visual information which, properly interpreted, can provide a different point of view to classical vision problems. Indeed, in the last decade, Brain Computer Interfaces (BCIs) and Wi-Fi signals have been used in a wide range of issues; now these technologies are robust enough to be used also for vision tasks (and not only).

In this presentation, an overview on Computer Vision and non-visual information supported by Artificial Intelligence for smart systems and emulation of human skills is presented. In addition, some new frontiers for BCIs and Wi-Fi signals for the “blind” vision are also reported.