12 Jul 2017
10:30
A105
P2: Software and Applications
| 1. Kinect as Master of Puppets: Animating Avatars for Virtual and Augmented Reality | ||
| Mateo Čobanov, Barbara Džaja and Josip Musić (University of Split, Croatia) | ||
| Today, Virtual Environment (VE) can be variated in two different,directions, Virtual Reality (VR) and Augmented Reality (AR). VR is,completely immersible to the subject on his way to the synthetic world,and one cannot perceive the real environment around him. On the other,hand, AR uses digital or computer generated information like images,,audio, video, integrating them in a real-time environment. AR,technically can be used to stimulate all five senses, but its most,common use was and still is the visual one. This paper examines,possibilities and challenges of using Microsoft Kinect sensor as tool,for animating characters in VR. Paper also explores AR created for,Android platform. Two case studies were performed, one on a desktop,computer (VR) and another on Android mobile phone (AR). Results are,presented and discussed as well as possible improvements. | ||
| 2. Development of Modular Unmanned Surface Vehicle for Research and Education | ||
| Josip Vasilj, Ivo Stančić, Josip Musić and Tamara Grujić (University of Split, Croatia) | ||
| Mobile robots are used for years as a valuable educational and research,tool in form of available open-platform designs and Do-It-Yourself kits.,Rapid development and costs reduction of Unmanned Air Vehicles (UAV),and ground based mobile robots in recent years allowed researchers and,students to utilize them as an affordable education platform. Despite of,recent developments in the area of ground and airborne robotics, only,few examples of Unmanned Surface Vehicle (USV) platforms targeted for,educational and research purposes can be found. Aim of this paper is to,present development of open-design USV drone with integrated multi-level,control hardware architecture. Catamaran type surface drone is planned,to be utilized as a research and education platform for students, where,various control algorithms, communication interfaces and sensor payloads,can be implemented and tested. Proposed USV enables autonomous,navigation or direct control over wireless radio link, separate,development of algorithms for optimal propulsion control, navigation and,communication with the ground-based control station. Whole design is,planned to be highly modular, where each component can be replaced or,modified according to desired task, payload or environmental conditions. | ||
| 3. Wearable RFID devices and cloud platform for efficient waste management | ||
| Cosimo Salvatore, Alfredo Salvatore and Angelo Primiani (Sensor ID, Italy) | ||
| Many municipalities are learning the importance of new technologies in,recycling industries, such as radio frequency identification (RFID), in,order to make more efficient the waste management system and to define a,real, precise and correct fee for citizens.,The introduction of tracking and control systems based on RFID,technology has the goal to increase the rate of differentiated material,and avoid bad behaviour of citizens thanks to the continuous monitoring,of collecting process. In this framework, Sensor ID has developed the SmartWasteID system based,on an RFID wearable reader, and a cloud platform that allow to automate,the entire garbage collection process. |
||
| 4. Human detection in aerial images gathered with Unmanned Aerial Vehicles (UAV) using Convolutional Neural Networks | ||
| Mirela Kundid Vasić (University of Mostar, Bosnia and Herzegovina); Vladan Papić (University of Split, Croatia) | ||
| In recent years Unmanned Aerial Vehicles (UAVs) are widely applied in,several areas of action, primarily due to several advantages such as,mobility, low weight and embedded image processing capabilities. In,search and rescue (SAR), one factor influencing the victim’s survival is,how quickly can that person be found. The problem of detecting people,in this case can be viewed as one of the tasks of processing and,examining aerial images. Testing many aerial photos is a handy process.,Automatic classification with the help of conventional machine learning,methods is a possible and also demanding task that gives promising,results for this problem, but there are a few different problems.,Conventional methods primarily requiring a set of predefined features,that need to be extracted from the objects we want to detect. The,problem with using conventional methods is that the objects in the,images are very small and defining a series of descriptive elements is,almost impossible. This paper tries to avoid such an approach and,considers the possibility of using the Convolutional Neural Networks to,learn possible features based on various convolution filters. Using the,DetectNet detection architecture through transfer learning from,pretrained models such as GoogLeNet, this paper considers approaches to,the application of these methods to achieve new results in the field of,image processing for human in distress detection. Success in the,implementation of a fast and efficient detection method for Unmanned,Aerial Vehicles would have resulted in easier finding of victims in,areas of low vegetation in the Mediterranean, such as southern,Herzegovina and Dalmatia. | ||
| 5. A Video Compression Method for Low Power Consumption Applications in Sensors Networks |
||
| Ben-Shung Chow (National Sun Yat-Sen University, Taiwan) | ||
| We proposed a video compression method for mobile video sensors networks by using binary video transmitted in low resolution. This visual quality is tolerable under the consideration of the convenient receiver and the limited bandwidth in communication. This compression method is based upon a new idea of shape compensation to replace the conventional residue Discrete Cosine Transform (DCT) coding. We develop an efficient shape compensation algorithm to a make this video compression suitable for activity recognition in sensor networks. |