The first hybridintegratedoptical phaselock loop (OPLL) for use in high spectral purity photonic terahertz sources was presented. We have achieved the necessary short loop delay to lock a 1MHz linewidth slave laser. The laser and photodetectors are InPbased and flip chip bonded to silicon daughter boards, which are in turn attached to the motherboard. Delay between the slave laser and photodiode was approximately 50 ps. The mixing product between slave and master sources has a linewidth of less than 1 kHz and achieved phase noise less than 80 dBc/Hz at an offset of 10 kHz. This integrated OPLL circuit was used with an optical comb source to generate high spectral purity signals at frequencies up to 300 GHz with linewidths less than 1 kHz.
COBISS.SI-ID: 8260948
A classical hierarchical menu structure of MS Word application was presented with multiple simultaneous sounds. In the user study with blind and visually impaired users we evaluated the efficiency of various numbers of simultaneous sounds as well as their spatial distribution (horizontal and vertical). The results demonstrated that the use of multiple simultaneous sounds increases the information flow between user and a computer and improves the navigation but decreases the speed of interaction due to very high cognitive workload.
COBISS.SI-ID: 8085332
In this paper we address two issues concerning real world time-continuous emotion detection from videos of users’ faces: (i) the impact of weak ground truth on the emotion detection accuracy and (ii) the impact of the users’ facial expressiveness on the emotion detection accuracy. We implemented an appearance based emotion detection algorithm that uses Gabor features and nearest neighbors classifier. We tested the performance of this algorithm on two datasets with different ground truth strengths (a firm ground truth dataset and a weak ground truth dataset). Then we split the dataset into three subsets reflecting different levels of users’ facial expressiveness (low, mid and high) and performed separate emotion detection.
COBISS.SI-ID: 9919060
State of the art content-based recommender systems usually rely on data-centric metadata (e.g. genre, actors ...). We introduced a set of user-centric metadata that describe the users' emotive responses. We compared the accuracy of recommended items in a content-based recommender system for images using (i) generic, data-centric, metadata and (ii) affective, user-centric, metadata. The results have shown that the recommender system with affective metadata yields significantly better accuracy than the same system with generic metadata.
COBISS.SI-ID: 7907412
In case of modern dynamic and interactive multimedia-rich services the importance of achieved Quality of Experience (QoE) equals or exceeds importance of Quality of Service (QoS). The paper presents results of detailed analysis on QoE and QoS profiling on all communication layers and presents modelling and simulations of QoE in NGN network using IMS, RACS and NASS subsystems. Presented models are based exclusively on existent profiles, functionalities, procedures and protocols available in an NGN.
COBISS.SI-ID: 7862612