University of Rijeka

About: University of Rijeka is a education organization based out in Rijeka, Croatia. It is known for research contribution in the topics: Population & Tourism. The organization has 3471 authors who have published 7993 publications receiving 110386 citations. The organization is also known as: Rijeka University & Sveučilište u Rijeci.

Speaker de-identification using diphone recognition and speech synthesis

Abstract: The paper addresses the problem of speaker (or voice) de-identification by presenting a novel approach for concealing the identity of speakers in their speech. The proposed technique first recognizes the input speech with a diphone recognition system and then transforms the obtained phonetic transcription into the speech of another speaker with a speech synthesis system. Due to the fact that a Diphone RecOgnition step and a sPeech SYnthesis step are used during the de-identification, we refer to the developed technique as DROPSY. With this approach the acoustical models of the recognition and synthesis modules are completely independent from each other, which ensures the highest level of input speaker de-identification. The proposed DROPSY-based de-identification approach is language dependent, text independent and capable of running in real-time due to the relatively simple computing methods used. When designing speaker de-identification technology two requirements are typically imposed on the de-identification techniques: i) it should not be possible to establish the identity of the speakers based on the de-identified speech, and ii) the processed speech should still sound natural and be intelligible. This paper, therefore, implements the proposed DROPSY-based approach with two different speech synthesis techniques (i.e, with the HMM-based and the diphone TD-PSOLA-based technique). The obtained de-identified speech is evaluated for intelligibility and evaluated in speaker verification experiments with a state-of-the-art (i-vector/PLDA) speaker recognition system. The comparison of both speech synthesis modules integrated in the proposed method reveals that both can efficiently de-identify the input speakers while still producing intelligible speech.

Ball Detection Using Yolo and Mask R-CNN

Abstract: Many computer vision applications rely on accurate and fast object detection, and in our case, ball detection serves as a prerequisite for action recognition in handball scenes. We compare the performance of two of the state-of-the-art convolutional neural network-based object detectors for the task of ball detection in non-staged, real-world conditions. The comparison is performed in terms of speed and accuracy measures on a dataset comprising custom handball footage and a sample of images obtained from the Internet. The performance of the models is compared with and without additional training with examples from our dataset.

Martensitic stainless steel AISI 420—mechanical properties, creep and fracture toughness

Abstract: In this paper some experimental results and analyses regarding the behavior of AISI 420 martensitic stainless steel under different environmental conditions are presented. That way, mechanical properties like ultimate tensile strength and 0.2 percent offset yield strength at lowered and elevated temperatures as well as short-time creep behavior for selected stress levels at selected elevated temperatures of mentioned material are shown. The temperature effect on mentioned mechanical properties is also presented. Fracture toughness was calculated on the basis of Charpy impact energy. Experimentally obtained results can be of importance for structure designers.

Molecular and physical characteristics of aerosol at a remote free troposphere site: implications for atmospheric aging

Abstract: . Aerosol properties are transformed by atmospheric processes during long-range transport and play a key role in the Earth's radiative balance. To understand the molecular and physical characteristics of free tropospheric aerosol, we studied samples collected at the Pico Mountain Observatory in the North Atlantic. The observatory is located in the marine free troposphere at 2225 m above sea level, on Pico Island in the Azores archipelago. The site is ideal for the study of long-range-transported free tropospheric aerosol with minimal local influence. Three aerosol samples with elevated organic carbon concentrations were selected for detailed analysis. FLEXPART retroplumes indicated that two of the samples were influenced by North American wildfire emissions transported in the free troposphere and one by North American outflow mainly transported within the marine boundary layer. Ultrahigh-resolution Fourier transform ion cyclotron resonance mass spectrometry was used to determine the detailed molecular composition of the samples. Thousands of molecular formulas were assigned to each of the individual samples. On average ∼60 % of the molecular formulas contained only carbon, hydrogen, and oxygen atoms (CHO), ∼30 % contained nitrogen (CHNO), and ∼10 % contained sulfur (CHOS). The molecular formula compositions of the two wildfire-influenced aerosol samples transported mainly in the free troposphere had relatively low average O∕C ratios ( 0.48±0.13 and 0.45±0.11 ) despite the 7–10 days of transport time according to FLEXPART. In contrast, the molecular composition of the North American outflow transported mainly in the boundary layer had a higher average O∕C ratio ( 0.57±0.17 ) with 3 days of transport time. To better understand the difference between free tropospheric transport and boundary layer transport, the meteorological conditions along the FLEXPART simulated transport pathways were extracted from the Global Forecast System analysis for the model grids. We used the extracted meteorological conditions and the observed molecular chemistry to predict the relative-humidity-dependent glass transition temperatures ( Tg ) of the aerosol components. Comparisons of the Tg to the ambient temperature indicated that a majority of the organic aerosol components transported in the free troposphere were more viscous and therefore less susceptible to oxidation than the organic aerosol components transported in the boundary layer. Although the number of observations is limited, the results suggest that biomass burning organic aerosol injected into the free troposphere is more persistent than organic aerosol in the boundary layer having broader implications for aerosol aging.