Personal Presentation
- 1. A Quick Personal Overview Antonio Di Maio Politecnico di Bari - Italy Last Update: 16 January 2016
- 2. Outline ¡ñ Studies and Curriculum ¡ñ Theses and Projects
- 3. My Studies and Curriculum Year Institution Title Final Mark 2016 Politecnico di Bari Master in Computer Science Engineering 110/110 Honors* 2013 Politecnico di Bari Bachelor in IT and Automation Engineering 110/110 Honors 2009 High School for IT Diploma of IT Technician and Programmer 100/100 Honors ¡ñ I have spent 6 months in the DFKI (Saarbr¨¹cken, Germany) writing my Master Thesis in Human- Machine Interaction during an Erasmus+ Programme ¡ñ During my studies I have developed a passion for ¡ð Computer Vision ¡ð Internet of Things ¡ð Computer Security ¡ð Artificial Intelligence ¡ð Embedded and Distributed Systems
- 4. Master Thesis: Design of an application for tracking a human target with a drone ¡ñ This project uses a classifier to start the tracking process and an experimental application of Optical Flow algorithm to track the target in a lightweight way ¡ñ The experimental nature of the tracking algorithm does not rely on Color Information ¡ñ The control feedback is sent to the engines using an experimental linear algorithm ¡ñ The project can be extended applying the results to a fleet of drones or to follow a multiple target
- 5. Master Thesis: Design of an application for tracking a human target with a drone ¡ñ Applications and Future Works ¡ð Personal Security ¡ö Target Tracking and Following (People, Vehicles¡) ¡ö Personal Assistant ¡ð Medical ¡ö Assistance for weak subjects ¡ö Falls detection in outdoor environment ¡ð Monitoring ¡ö Mobile traffic state cameras ¡ð Human-Machine Interaction ¡ö Controlling a system with body and hand gestures ¡ö Teaching a drone to do an inconvenient task for humans
- 6. Bachelor Thesis: Design of an Android application for Eye Blink Detection ¡ñ This project uses a set of Viola-Jones classifiers to find the position of the face and eyes ¡ñ The blink detection is performed through a large set of experimental methods, involving Integral Projection Functions, Hu Moments, Hough Transforms, Background Subtraction, Color segmentation and many more. ¡ñ This project represents my current job: I am developing new features as a freelance project for an asian company
- 7. Bachelor Thesis: Design of an Android application for Eye Blink Detection ¡ñ Applications and Future Works ¡ð Vehicle Security ¡ö Aliveness Check ¡ö Driver Alertness State Monitoring ¡ð Medical ¡ö Early Diagnosis for eye blink related diseases ¡ö Stress Monitoring ¡ð Human-Machine Interaction ¡ö Controlling a system with eye blinks and eye gaze ¡ö Communicating without voice or hand gestures
- 8. Medical Project and Scientific Publication ¡ñ ¡°A novel approach to evaluate blood parameters using computer vision techniques¡± ¡ð The study investigates whether the characteristics of digital images of eye conjunctivas are correlated with blood parameters, in order to predict their numeric values or warn the patient if they go under a safety threshold ¡ð This can avoid painful, expensive and time-consuming blood samplings ¡ð Our team has designed a physical device to standardize the process of image acquisition, based on embedded systems (RaspberryPi) ¡ð An automatic segmentation algorithm has been developed to avoid the manual palpebral segmentation by the medical expert ¡ð In order to prove that there is a (linear) correlation, a statistical analysis phase has been carried out, calculating the Pearson Correlation Indexes and training Support Vector Machines
- 9. IoT Project: A WSN Mobile Navigator ¡ñ TLSensing Mobile ¡ð It can browse the information content of a WSN using the OM2M application paradigm ¡ð It provides both general network statistics or navigation of a particular node¡¯s data, with graphs ¡ð The sink is connected to an embedded system (PandaBoard) that runs the application server that manages the requests either over HTTP or CoAP
- 10. Other Projects ¡ñ ArduLock - An Android-Arduino Powered Security system ¡ñ Cardiac arrhythmias detection using the Wavelet Transform ¡ñ Java-Python translator, using CUP and JFlex ¡ñ Small household projects to solve little issues (Arduino, RPi, Chromecast) ¡ñ ....
- 11. Award: Google Hackathon 2014 First Prize ¡ñ Healthwear ¡ð It uses a smartwatch to send help requests through a smartphone ¡ð The smartwatch¡¯s accelerometer can detect when a person falls and send several useful health parameters such as heart rate and GPS position ¡ð The operator visualizes the GPS position on a map and can send a specialized team to rescue the patient. It is also possible to set the App to send an SMS to or call a friend's number.
- 12. Thanks for your interest :)