AIML

AIML

[VPH] 

Visual Perception in HRI

PhD program in Engineering in Computer Science - A.Y. 2020/21 (fall semester) - 1.5 ECTS/CFU (type B)

ONLINE

Abstract

As robotics evolves towards application fields in which humans cooperate with robots, working closer and closer, the requirements for human robot interactions increase. Visual perception is an important component for human–robot interaction processes in robotic systems. Interaction between humans and robots depends on the reliability of the robotic vision systems. The analysis of activities, motions, skills, and behaviors of humans and robots are generally addressed by using the features of a moving human body (or body part). The human motion behavior is then analyzed by body movement kinematics, and the trajectory of the target is used to identify the objects and the human target. The process of human target identification and gesture recognition in a quite non-trivial problem. In this series of lectures we will focus on the context of Human-Robot Interaction (HRI) along with the related problems on the field of vision and perception, applied to robotic systems. We will devise the typical characteristics of vision and perception related hardware device, as well as the relative software systems and solutions. We will explore the known approaches characterizing well known visual recognition systems, as well as the most important algorithmic solutions for people targeting and body parts recognition. A theoretical and practical framework will be given with several example. Finally we will discuss the state of the art on human-centric vision analysis and explain the importance of the matter relatively to human-based interfaces of computer/robots with special interest in human motion and activity recognition. We will also devise several tracking systems and motion oriented context and object recognition techniques, with emphasis on deep learning techniques applied to visual recognition. Finally we will compare the applicability of such techniques to human motion classification and the related application on the field of Human-Robot Interaction.

Exam (1.5 ECTS/CFU credits - type B)

The exam consists of a development project using techniques related to the course topics. The candidates will be required to deliver the developed code, any correlated datasets, and a report. The report will be structured as a scientific publication and must include a description of the problem, the solution developed and the results obtained. Passing the exam entitles the candidate with 1.5 ECTS/CFU of type B, as specified in the regulations of the doctoral program.
  

ONLINE DIDACTIC

INSTRUCTIONS

1. Go to https://classroom.google.com/

2. subscribe to the course with the following code: fnzc6sr

3. follow the instructions given into the virtual classroom

4. use the google meet link you will find in the classroom header

Thank you for your cooperation.