Goran Martinović

Emotion Recognition System by a Neural Network Based Facial Expression Analysis

Human-computer interfaces are getting more complex every day with the purpose of easing the use of computers and enhancing the overall user experience. Since research has shown that a majority of human interaction comes from non-verbal communication, user emotion detection is one of the directions that can be taken to enhance the overall user experience. This paper proposes a system for human emotion recognition by analyzing key facial regions using principal component analysis and neural networks. The proposed system has been trained and tested on the FEEDTUM database where it achieved a relatively high average score of correct recognition and therefore showed promise for future development.

Single-Commodity Vehicle Routing Problem with Pickup and Delivery Service

We present a novel variation of the vehicle routing problem (VRP). Single commodity cargo with pickup and delivery service is considered. Customers are labeled as either cargo sink or cargo source, depending on their pickup or delivery demand. This problem is called a single commodity vehicle routing problem with pickup and delivery service (1-VRPPD). 1-VRPPD deals with multiple vehicles and is the same as the single-commodity traveling salesman problem (1-PDTSP) when the number of vehicles is equal to 1. Since 1-VRPPD specializes VRP, it is 𝒩𝒫 hard in the strong sense. Iterative modified simulated annealing (IMSA) is presented along with greedy random-based initial solution algorithm. IMSA provides a good approximation to the global optimum in a large search space. Experiment is done for the instances with different number of customers and their demands. With respect to average values of IMSA execution times, proposed method is appropriate for practical applications.

A differential evolution approach to dimensionality reduction for classification needs

Abstract The feature selection problem often occurs in pattern recognition and, more specifically, classification. Although these patterns could contain a large number of features, some of them could prove to be irrelevant, redundant or even detrimental to classification accuracy. Thus, it is important to remove these kinds of features, which in turn leads to problem dimensionality reduction and could eventually improve the classification accuracy. In this paper an approach to dimensionality reduction based on differential evolution which represents a wrapper and explores the solution space is presented. The solutions, subsets of the whole feature set, are evaluated using the k-nearest neighbour algorithm. High quality solutions found during execution of the differential evolution fill the archive. A final solution is obtained by conducting k-fold crossvalidation on the archive solutions and selecting the best one. Experimental analysis is conducted on several standard test sets. The classification accuracy of the k-nearest neighbour algorithm using the full feature set and the accuracy of the same algorithm using only the subset provided by the proposed approach and some other optimization algorithms which were used as wrappers are compared. The analysis shows that the proposed approach successfully determines good feature subsets which may increase the classification accuracy.

PVCM: Assisting multi-hop communication in vehicular networks using parked vehicles

Multi-hop communication in vehicular ad hoc networks is a challenging task due to the dynamic and unbalanced nature of these networks. In the suburban areas or during the night when the density of vehicles is low, multi-hop communication is only possible if (i) the vehicles store-carry-forward the messages or (ii) by using roadside units. Each of the solutions has its drawback. The former introduces delivery delay and the latter is an expensive solution. In this paper, we propose the Parked Vehicle Communication Model (PVCM) to use the parked vehicles along the roads to act as relay nodes and assist the communication in low density situations. We argue that the number and the distribution of parked vehicles is suitable to serve our purpose. We also show that the automotive battery can be utilized to keep the on-board unit running without draining all its energy.

Web based distant monitoring and control for greenhouse systems using the Sun SPOT modules

This article presents a solution for distant monitoring and control of the greenhouse system via internet. It has a web based user interface with possibility of managing the sensors and actuators. Also allows to log events on the system like humidity, temperature, light etc. This Web based distant monitoring system uses a high end PIC microcontroller based developer board with Ethernet adapter that will allow you to monitor and control the laboratory from any Internet connection. This solution offers a live video stream from the greenhouses camera and allows monitoring the growth of the plants. With the usage of Sun SPOT based wireless sensor network is realized the environment monitoring. The described system is highly customizable and easily adaptable to particular deployment requirements.

Towards an energy efficient SoC computing cluster

High performance computing (HPC) systems are omnipresent in processing huge amounts of data. However, power consumption of these systems is becoming far from negligible, forcing software developers and hardware providers to deviate from usual performance plans, and to address power consumption and costs. As a result, small-scale applications are becoming infeasible to run on those systems, making room for the power-efficient HW newcomers, such as SoC (System on Chip), to take that role. The idea of composing a large number of small SoCs in a computing cluster can result in a powerful, green, and a simple platform capable of delivering enough performance for the small applications while maintaining low energy consumption and costs. In this paper we present a system composed of four LAN-interconnected dual-core ARM based SoC devices, powered by a single power supply. We investigate applicability of such a system by monitoring its performance and energy efficiency and stability while using common HPC paradigms: message passing and thread programming.

Quality of Service in Wireless Sensor Networks: A Survey and Related Patents

Rapid advances in Wireless Sensor Networks (WSNs) indicate that they are becoming increasingly complex. Consequently users and applications are becoming more demanding. Due to unique characteristics of WSNs, like small dimensions and limited resources and capabilities, Quality of Service (QoS) is imposed as one of the key factors of WSNs. In this paper, we surveyed two main approaches for QoS provisioning in WSNs: layered and cross-layer approach. QoS provisioning with layered approach is surveyed in three WSN layers: MAC, network and transport layer. Current developments show that they can be efficiently used for QoS provisioning. However, they consider QoS only as layer specific isolated set of problems and they are highly dependent on the performance of other layers. Cross-layer approach does not have the restrictions as layered approach and hence can dispose with information from all layers of the communication protocol stack. Although it has huge potential to become the most efficient solution for QoS provisioning in WSNs, current development indicate that there are still many issues and challenges that need to be overcome. Since the concept of the QoS is relatively new in WSNs, there are not a large number of patents currently dealing with this issue, however but in coming years a large increase in the number of such patents is expected. Available patents in this domain are described in the paper.

Spatial Masking and Perceived Video Quality in Multimedia Applications

In modern communication systems with video transmission as their important part, objective video quality evaluation is necessary in the QoS (quality of service) assessment. This article analyzes spatial masking effects on the visibility of compression artifacts, and consequently its importance in the digital video quality evaluation process. Luminance masking and texture masking are mechanisms referred to as having a great influence on the objective quality metrics fidelity. Eleven maskers (6 for luminance and 5 for texture) are constructed, and their effectiveness is analyzed first for each masker alone, and then for combinations of luminance and texture maskers. By nonlinear addition of luminance and texture maskers used in a simple pixel-based quality metric a high correlation with subjective mean opinion scores (MOS) is obtained (Pearson correlation coefficient higher than 0.95).

Undergraduate teaching of real-time scheduling algorithms by developed software tool

Theoretical foundations of scheduling in the course Real-Time Systems of the undergraduate studies program are extended by the usage of the developed visual software tool. The tool enables teaching of simulated fixed-priority scheduling algorithms. The user simply sets the simulation and algorithm properties and loads the prepared input task sets from the database. Results are obtained within the program or from external applications. The paper deals with the problem of scheduling the mixture of periodic and aperiodic tasks in a uniprocessor system. Results are displayed in the form of a flow report, table report, and diagrams of dependencies of evaluation parameters on the change of periodic and aperiodic workload. The flow report explains the algorithm behavior and the table form and dependency diagrams enable a more precise scheduling analysis and mutual comparison, respectively. Some of the evaluation parameters are the following: aperiodic response time, number of tasks with missed deadlines, utilization of resource and server tasks, resource-breakdown utilization. The software tool improves teaching of the course and the final success of students regarding relevant exam exercises.

A Self Manageable Rule Driven Enterprise Application

In this paper we present a design and a prototype of a self-manageable application. Self management is accomplished by usage of a rule based autonomic manager. The application demonstrates principles for combining enterprise servers and custom developed components into a system with self-configuring, self-healing and self-optimizing characteristics.