Search

Applications often require a fast, single-threaded search algorithm over sorted data, typical in table-lookup operations. We explore various search algorithms for a large number of search candidates over a relatively small array of logarithmically-distributed sorted data. These include an innovative hash-based search that takes advantage of floating point representation to bin data by the exponent. Algorithms that can be optimized to take advantage of SIMD vector instructions are of particular interest. We then conduct a case study applying our results and analyzing algorithmic performance with the EOSPAC package. EOSPAC is a table look-up library for manipulation and interpolation of SESAME equation-of-state data. Our investigation results in a couple of algorithms with better performance with a best case eight times speedup over the original EOSPAC Hunt-and-Locate implementation. Our techniques should generalize to other instances of search algorithms seeking to get a performance boost from vectorization.

Quality control of products plays an important role in various stages of the manufacturing process. In particular the final control of the quality of a product before being shipped to a customer is crucial for maintaining customer satisfaction and avoiding costly recalls. Automating quality inspection and integrating it into a seamless Industry 4.0 setting is therefore an important topic in factory automation.We present early work towards an automated product inspection station. Our inspection station features a 3D scanner as well as a Virtual Reality headset for remote human inspection. In addition, our concept provides for automated analysis of scans in the cloud. We present an architectural concept as well as an early prototype.

The characterization of solar surface differential rotation (SDR) from disk-integrated chromospheric measurements has important implications for the study of differential rotation and dynamo processes in other stars. Some chromospheric lines, such as Ca II K, are very sensitive to the presence of activity on the disk and are an ideal choice for investigating SDR in Sun-as-a star observations. Past studies indicate that when the activity is low, the determination of Sun's differential rotation from integrated-sunlight measurements becomes uncertain. However, our study shows that using the proper technique, SDR can be detected from these type of measurements even during periods of extended solar minima. This paper describes results from the analysis of the temporal variations of Ca II K line profiles observed by the Integrated Sunlight Spectrometer (ISS) during the declining phase of Cycle 23 and the rising phase of Cycle 24, and discusses the signature of SDR in the power spectra computed from time series of parameters derived from these profiles. The described methodology is quite general, and could be applied to photometric time series of other Main-Sequence stars for detecting differential rotation., Accepted for publication in The Astrophysical Journal. 38 pages, 10 figures

We compare photospheric line-of-sight magnetograms from the Synoptic Long-term Investigations of the Sun (SOLIS) vector spectromagnetograph (VSM) instrument with observations from the 150-foot Solar Tower at Mt. Wilson (MWO), Helioseismic and Magnetic Imager (HMI) on Solar Dynamics Observatory (SDO), and Michelson Doppler Imager (MDI) on Solar and Heliospheric Observatory (SOHO). We find very good agreement between VSM and the other data sources for both disk-averaged flux densities and pixel-by-pixel measurements. We show that the VSM mean flux density time series is of consistently high signal-to-noise with no significant zero-offsets. We discuss in detail some of the factors -spatial resolution, flux dependence and position on the solar disk- affecting the determination of scaling between VSM and SOHO/MDI or SDO/HMI magnetograms. The VSM flux densities agree well with spatially smoothed data from MDI and HMI, although the scaling factors show clear dependence on flux density. The factor to convert VSM to HMI increases with increasing flux density (from $\approx$1 to $\approx$1.5). The nonlinearity is smaller for the VSM vs. ~SOHO/MDI scaling factor (from $\approx$1 to $\approx$1.2)., Accepted for publication in Solar Physics

We present a methodology to determine the best turbulence closure for an eddy-permitting ocean model through measurement of the error-landscape of the closure's subgrid spectral transfers and flux. We apply this method to 6 different closures for forced-dissipative simulations of the barotropic vorticity equation on a f-plane (2D Navier-Stokes equation). Using a high-resolution benchmark, we compare each closure's model of energy and enstrophy transfer to the actual transfer observed in the benchmark run. The error-landscape norms enable us to both make objective comparisons between the closures and to optimize each closure's free parameter for a fair comparison. The hyper-viscous closure most closely reproduces the enstrophy cascade, especially at larger scales due to the concentration of its dissipative effects to the very smallest scales. The viscous and Leith closures perform nearly as well, especially at smaller scales where all three models were dissipative. The Smagorinsky closure dissipates enstrophy at the wrong scales. The anticipated potential vorticity closure was the only model to reproduce the upscale transfer of kinetic energy from the unresolved scales, but would require high-order Laplacian corrections in order to concentrate dissipation at the smallest scales. The Lagrangian-averaged alpha-model closure did not perform successfully for forced 2D isotropic Navier-Stokes: small-scale filamentation is only slightly reduced by the model while small-scale roll-up is prevented. Together, this reduces the effects of diffusion., 44 pages, 21 figures, 1 Appendix, submitted to Ocean Modeling

Understanding the in situ amplification of large scale magnetic fields in turbulent astrophysical rotators has been a core subject of dynamo theory. When turbulent velocities are helical, large scale dynamos that substantially amplify fields on scales that exceed the turbulent forcing scale arise, but the minimum sufficient fractional kinetic helicity f_h,C has not been previously well quantified. Using direct numerical simulations for a simple helical dynamo, we show that f_h,C decreases as the ratio of forcing to large scale wave numbers k_F/k_min increases. From the condition that a large scale helical dynamo must overcome the backreaction from any non-helical field on the large scales, we develop a theory that can explain the simulations. For k_F/k_min>8 we find f_h,C< 3%, implying that very small helicity fractions strongly influence magnetic spectra for even moderate scale separation., 5 pages, 4 figures

The magnetic field plays a pivotal role in many fields of Astrophysics. This is especially true for the physics of the solar atmosphere. Measuring the magnetic field in the upper solar atmosphere is crucial to understand the nature of the underlying physical processes that drive the violent dynamics of the solar corona -- that can also affect life on Earth. SolmeX, a fully equipped solar space observatory for remote-sensing observations, will provide the first comprehensive measurements of the strength and direction of the magnetic field in the upper solar atmosphere. The mission consists of two spacecraft, one carrying the instruments, and another one in formation flight at a distance of about 200m carrying the occulter to provide an artificial total solar eclipse. This will ensure high-quality coronagraphic observations above the solar limb. Solmex integrates two spectro-polarimetric coronagraphs for off-limb observations, one in the EUV and one in the IR, and three instruments for observations on the disk. The latter comprises one imaging polarimeter in the EUV for coronal studies, a spectro-polarimeter in the EUV to investigate the low corona, and an imaging spectro-polarimeter in the UV for chromospheric studies. SOHO and other existing missions have investigated the emission of the upper atmosphere in detail (not considering polarization), and as this will be the case also for missions planned for the near future. Therefore it is timely that SolmeX provides the final piece of the observational quest by measuring the magnetic field in the upper atmosphere through polarimetric observations.

We present high spatial and temporal resolution Ca II 8542 observations of a kink wave in an on-disk chromospheric active region fibril. The properties of the wave are similar to those observed in off-limb spicules. From the observed phase and period of the wave we determine a lower limit for the field strength in the chromospheric active region fibril located at the edge of a sunspot to be a few hundred Gauss. We find indications that the event was triggered by a small-scale reconnection event higher up in the atmosphere., 16 pages, 6 figures. Accepted for publication in ApJ

With the help of a model of magnetohydrodynamic (MHD) turbulence tested previously, we explore high Reynolds number regimes up to equivalent resolutions of 6000^3 grid points in the absence of forcing and with no imposed uniform magnetic field. For the given initial condition chosen here, with equal kinetic and magnetic energy, the flow ends up being dominated by the magnetic field, and the dynamics leads to an isotropic Iroshnikov-Kraichnan energy spectrum. However, the locally anisotropic magnetic field fluctuations perpendicular to the local mean field follow a Kolmogorov law. We find that the ratio of the eddy turnover time to the Alfven time increases with wavenumber, contrary to the so-called critical balance hypothesis. Residual energy and helicity spectra are also considered; the role played by the conservation of magnetic helicity is studied, and scaling laws are found for the magnetic helicity and residual helicity spectra. We put these results in the context of the dynamics of a globally isotropic MHD flow which is locally anisotropic because of the influence of the strong large-scale magnetic field, leading to a partial equilibration between kinetic and magnetic modes for the energy and the helicity., 11 pages, 7 figures, submitted to Phys. Rev. E

We study the velocity structure of penumbral filaments in the deep photosphere to obtain direct evidence for the convective nature of sunspot penumbrae. A sunspot was observed at high spatial resolution with the 1-m Swedish Solar Telescope in the deep photospheric C I 5380 {\AA} absorption line. The Multi-Object Multi-Frame Blind Deconvolution (MOMFBD) method is used for image restoration and straylight is filtered out. We report here the discovery of clear redshifts in the C I 5380 {\AA} line at multiple locations in sunspot penumbral filaments. For example, bright head of filaments show larger concentrated blueshift and are surrounded by darker, redshifted regions, suggestive of overturning convection. Elongated downflow lanes are also located beside bright penumbral fibrils. Our results provide the strongest evidence yet for the presence of overturning convection in penumbral filaments and highlight the need to observe the deepest layers of the penumbra in order to uncover the energy transport processes taking place there., Comment: accepted for ApJL

One goal of helioseismology is to determine the subsurface structure of sunspots. In order to do so, it is important to understand first the near-surface effects of sunspots on solar waves, which are dominant. Here we construct simplified, cylindrically-symmetric sunspot models, which are designed to capture the magnetic and thermodynamics effects coming from about 500 km below the quiet-Sun $\tau_{5000}=1$ level to the lower chromosphere. We use a combination of existing semi-empirical models of sunspot thermodynamic structure (density, temperature, pressure): the umbral model of Maltby et al. (1986) and the penumbral model of Ding and Fang (1989). The OPAL equation of state tables are used to derive the sound speed profile. We smoothly merge the near-surface properties to the quiet-Sun values about 1mm below the surface. The umbral and penumbral radii are free parameters. The magnetic field is added to the thermodynamic structure, without requiring magnetostatic equilibrium. The vertical component of the magnetic field is assumed to have a Gaussian horizontal profile, with a maximum surface field strength fixed by surface observations. The full magnetic field vector is solenoidal and determined by the on-axis vertical field, which, at the surface, is chosen such that the field inclination is 45$^\circ$ at the umbral-penumbral boundary. We construct a particular sunspot model based on SOHO/MDI observations of the sunspot in active region NOAA 9787. The helioseismic signature of the model sunspot is studied using numerical simulations of the propagation of f, p$_1$, and p$_2$ wave packets. These simulations are compared against cross-covariances of the observed wave field. We find that the sunspot model gives a helioseismic signature that is similar to the observations., Comment: 18 pages, 9 Figures submitted to Solar Physics

We present the first observations of the transverse component of photospheric magnetic field acquired by the imaging magnetograph Sunrise/IMaX. Using an automated detection method, we obtain statistical properties of 4536 features with significant linear polarization signal. Their rate of occurrence is 1-2 orders of magnitude larger than values reported by previous studies. We show that these features have no characteristic size or lifetime. They appear preferentially at granule boundaries with most of them being caught in downflow lanes at some point in their evolution. Only a small percentage are entirely and constantly embedded in upflows (16%) or downflows (8%)., Accepted for the Sunrise Special Issue of ApJL

Context: Using unique quadrature observations with the two STEREO spacecraft, we investigate coronal dimmings at the onset of small-scale eruptions. In CMEs they are believed to indicate the opening up of the coronal magnetic fields at the start of the eruption. Aims: It is to determine whether coronal dimming seen in small-scale eruptions starts before or after chromospheric plasma ejection. Methods: One STEREO spacecraft obtained high cadence, 75 s, images in the He II 304A channel, and the other simultaneous images in the Fe IX/FeX 171A channel. We concentrate on two well-positioned chromospheric eruptions that occurred at disk center in the 171A images, and on the limb in 304A. One was in the quiet Sun and the other was in an equatorial coronal hole. We compare the timing of chromospheric eruption seen in the 304A limb images with the brightenings and dimmings seen on disk in the 171A images. Further we use off-limb images of the low frequency 171A power to infer the coronal structure near the eruptions. Results: In both the quiet Sun and the coronal hole eruption, on disk 171A dimming was seen before the chromospheric eruption, and in both cases it extends beyond the site of the chromospheric eruption. The quiet Sun eruption occurred on the outer edge of the enclosing magnetic field of a prominence and may be related to a small disruption of the prominence just before the 171A dimming. Conclusions: These small-scale chromospheric eruptions started with a dimming in coronal emission just like their larger counterparts. We therefore suggest that a fundamental step in triggering them was the removal of overlying coronal field., 4 pages, 8 figures. To appear A&A Letters. Movies accompanying this Letter are at http://www.mps.mpg.de/data/outgoing/innes/dims/

The subfilter-scale (SFS) physics of regularization models are investigated to understand the regularizations' performance as SFS models. The strong suppression of spectrally local SFS interactions and the conservation of small-scale circulation in the Lagrangian-averaged Navier-Stokes alpha-model (LANS-alpha) is found to lead to the formation of rigid bodies. These contaminate the superfilter-scale energy spectrum with a scaling that approaches k^1 as the SFS spectra is resolved. The Clark-alpha and Leray-alpha models, truncations of LANS-alpha, do not conserve small-scale circulation and do not develop rigid bodies. LANS-alpha, however, is closest to Navier-Stokes in intermittency properties. All three models are found to be stable at high Reynolds number. Differences between L^2 and H^1 norm models are clarified. For magnetohydrodynamics (MHD), the presence of the Lorentz force as a source (or sink) for circulation and as a facilitator of both spectrally nonlocal large to small scale interactions as well as local SFS interactions prevents the formation of rigid bodies in Lagrangian-averaged MHD (LAMHD-alpha). We find LAMHD-alpha performs well as a predictor of superfilter-scale energy spectra and of intermittent current sheets at high Reynolds numbers. We expect it may prove to be a generally applicable MHD-LES., Proceedings of the Quality and Reliability of Large-Eddy Simulations II 2009 Workshop, Springer ERCOFTAC volume 16 (2011); submitted to Journal of Scientific Computing; 22 pages, 8 figures

We demonstrate that, for the case of quasi-equipartition between the velocity and the magnetic field, the Lagrangian-averaged magnetohydrodynamics alpha-model (LAMHD) reproduces well both the large-scale and small-scale properties of turbulent flows; in particular, it displays no increased (super-filter) bottleneck effect with its ensuing enhanced energy spectrum at the onset of the sub-filter-scales. This is in contrast to the case of the neutral fluid in which the Lagrangian-averaged Navier-Stokes alpha-model is somewhat limited in its applications because of the formation of spatial regions with no internal degrees of freedom and subsequent contamination of super-filter-scale spectral properties. No such regions are found in LAMHD, making this method capable of large reductions in required numerical degrees of freedom; specifically, we find a reduction factor of 200 when compared to a direct numerical simulation on a large grid of 1536^3 points at the same Reynolds number., 22 pages, 9 figures; accepted Phys.Rev.E

We present two models for turbulent flows with periodic boundary conditions and with either rotation, or a magnetic field in the magnetohydrodynamics (MHD) limit. One model, based on Lagrangian averaging, can be viewed as an invariant-preserving filter, whereas the other model, based on spectral closures, generalizes the concepts of eddy viscosity and eddy noise. These models, when used separately or in conjunction, may lead to substantial savings for modeling high Reynolds number flows when checked against high resolution direct numerical simulations (DNS), the examples given here being run on grids of up to 1536^3 points., 8 pages, 4 figures

We determine how the differences in the treatment of the subfilter-scale physics affect the properties of the flow for three closely related regularizations of Navier-Stokes. The consequences on the applicability of the regularizations as SGS models are also shown by examining their effects on superfilter-scale properties. Numerical solutions of the Clark-alpha model are compared to two previously employed regularizations, LANS-alpha and Leray-alpha (at Re ~ 3300, Taylor Re ~ 790) and to a DNS. We derive the Karman-Howarth equation for both the Clark-alpha and Leray-alpha models. We confirm one of two possible scalings resulting from this equation for Clark as well as its associated k^(-1) energy spectrum. At sub-filter scales, Clark-alpha possesses similar total dissipation and characteristic time to reach a statistical turbulent steady-state as Navier-Stokes, but exhibits greater intermittency. As a SGS model, Clark reproduces the energy spectrum and intermittency properties of the DNS. For the Leray model, increasing the filter width decreases the nonlinearity and the effective Re is substantially decreased. Even for the smallest value of alpha studied, Leray-alpha was inadequate as a SGS model. The LANS energy spectrum k^1, consistent with its so-called "rigid bodies," precludes a reproduction of the large-scale energy spectrum of the DNS at high Re while achieving a large reduction in resolution. However, that this same feature reduces its intermittency compared to Clark-alpha (which shares a similar Karman-Howarth equation). Clark is found to be the best approximation for reproducing the total dissipation rate and the energy spectrum at scales larger than alpha, whereas high-order intermittency properties for larger values of alpha are best reproduced by LANS-alpha., Comment: 21 pages, 8 figures

Odour emissions from passenger ships were noticed while making field investigation for ship yard in Finland. Local authors wanted to know odour load of ships because they have planned to build a new residential area near to passenger harbour. Odour load from ships were determined with field investigation, emission measurements and dispersion calculation.

We study non-LTE inversions of the Ca II infrared triplet lines as a tool for inferring physical properties of the quiet Sun. The inversion code is successful in recovering the temperature, velocity and longitudinal magnetic flux density in the photosphere and chromosphere, but the height range where the inversions are sensitive is limited, especially in the chromosphere. We present results of inverting spectropolarimetric observations of the lines in a quiet Sun region. We find three distinct ranges in chromospheric temperature: low temperatures in the internetwork, high temperatures in the enhanced magnetic network and intermediate temperatures associated with low magnetic flux regions in the network. The differences between these regions become more pronounced with height as the plasma-$\beta$ decreases. These inversions support the picture of the chromosphere, especially close to the magnetic network, being highly inhomogeneous both in the vertical and horizontal directions., Comment: 31 pages, 13 figures, accepted to ApJ

The Ca II infrared triplet is one of the few magnetically sensitive chromospheric lines available for ground-based observations. We present spectropolarimetric observations of the 8498 A and 8542 A lines in a quiet Sun region near a decaying active region and compare the results with a simulation of the lines in a high plasma-beta regime. Cluster analysis of Stokes V profile pairs shows that the two lines, despite arguably being formed fairly close, often do not have similar shapes. In the network, the local magnetic topology is more important in determining the shapes of the Stokes V profiles than the phase of the wave, contrary to what our simulations show. We also find that Stokes V asymmetries are very common in the network, and the histograms of the observed amplitude and area asymmetries differ significantly from the simulation. Both the network and internetwork show oscillatory behavior in the Ca II lines. It is stronger in the network, where shocking waves, similar to those in the high-beta simulation, are seen and large self-reversals in the intensity profiles are common., 23 pages, 17 figures, accepted to ApJ some figures are low-res, for high-res email annap@ucar.edu

Users expect efficient ways of handling their personal audio and image data using only the mobile device. Data sharing with friends and interest groups is also becoming increasingly important. Browsing and manually searching through your own and a shared data library can be difficult. For example, the simplest traditional attributes like time and date are normally unknown for other people's data. In this paper a search application running on mobile device is described. The application relies on automatic content analysis and metadata generation for audio and image data. Two search paradigms - class based search and example based search - are supported. In mobile devices the small display size makes it difficult to handle large amounts of information quickly. For this reason efficient metadata visualization is important. In our system the content metadata is converted into visual form that is fast to browse through at one glance.

We present an extension of the Kármán-Howarth theorem to the Lagrangian averaged magnetohydrodynamic (LAMHD-α) equations. The scaling laws resulting as a corollary of this theorem are studied in numerical simulations, as well as the scaling of the longitudinal structure function exponents indicative of intermittency. Numerical simulations for a magnetic Prandtl number equal to unity are presented both for freely decaying and for forced two-dimensional magnetohydrodynamic (MHD) turbulence, solving the MHD equations directly, and employing the LAMHD-α equations at 1∕2 and 1∕4 resolution. Linear scaling of the third-order structure function with length is observed. The LAMHD-α equations also capture the anomalous scaling of the longitudinal structure function exponents up to order 8.

We present a video management system comprising a video server and a mobile camera-phone application called MobiCon, which allows users to capture videos, annotate them with metadata, specify digital rights management (DRM) settings, upload the videos over the cellular network, and share them with others. Once stored in the video server, users can then search their personal video collection via a web interface, and watch the video clips using a wide range of terminals. We describe the MobiCon architecture, compare it with related work, provide an overview of the video server, and illustrate a typical user scenario from the point of capture to video sharing and video searching. Our work takes steps forward in advancing the mobile camera-phone from a video playback device to a video production tool. We summarize field trial results conducted in the area of Oulu, Finland, which demonstrate that users can master the application quickly, but are unwilling to perform extensive manual annotations. Based on the user trial results and our own experience, we present future development directions for MobiCon, in particular, and the video management architecture, in general.

Although many different types of technologies for information systems have evolved over the last decades (such as databases, video systems, the Internet and mobile telecommunication), the integration of these technologies is just in its infancy and has the potential to introduce ”intelligent” systems. The CANDELA project, which is part of the European ITEA program, focuses on the integration of video content analysis in combination with networked delivery and storage technologies. To unleash the full potential of such integration, adaptive video-content analysis and retrieval techniques are being explored by developing several pilot applications.

We present an extension of the Karman-Howarth theorem to the Lagrangian averaged magnetohydrodynamic (LAMHD-alpha) equations. The scaling laws resulting as a corollary of this theorem are studied in numerical simulations, as well as the scaling of the longitudinal structure function exponents indicative of intermittency. Numerical simulations for a magnetic Prandtl number equal to unity are presented both for freely decaying and for forced two dimensional MHD turbulence, solving directly the MHD equations, and employing the LAMHD-alpha equations at 1/2 and 1/4 resolution. Linear scaling of the third-order structure function with length is observed. The LAMHD-alpha equations also capture the anomalous scaling of the longitudinal structure function exponents up to order 8., 34 pages, 7 figures author institution addresses added magnetic Prandtl number stated clearly

We present direct numerical simulations and Lagrangian averaged (also known as alpha-model) simulations of forced and free decaying magnetohydrodynamic turbulence in two dimensions. The statistics of sign cancellations of the current at small scales is studied using both the cancellation exponent and the fractal dimension of the structures. The alpha-model is found to have the same scaling behavior between positive and negative contributions as the direct numerical simulations. The alpha-model is also able to reproduce the time evolution of these quantities in free decaying turbulence. At large Reynolds numbers, an independence of the cancellation exponent with the Reynolds numbers is observed., Finite size box effects have been taken into account in the definition of the partition function. This has resulted in a more clear scaling in all figures. Several points are clarified in the text

Video management research has largely been ignoring the increased attractiveness of using camera-equipped mobile phones for the production of short home video clips, mostly considering them as additional channels for video consumption. The CANDELA project, which is part of the European ITEA program, focuses on the integration of video content analysis with advanced retrieval, mobile, networked delivery, and distributed storage technologies. In this paper, we present the CANDELA personal mobile multimedia management platform, which implements an end-to-end system for personal video production, retrieval, and consumption utilizing mobile devices and distributed databases.

The latest camera-equipped mobile phones and faster cellular networks have increased the interest in mobile multimedia services. But for content consumption, delivery and creation, the limited capabilities of mobile terminals require special attention. This paper introduces the Candela platform, an infrastructure that allows the creation, storage and retrieval of home videos with special consideration of mobile terminals. Candela features a J2ME-based video recording and annotation tool which permits the creation and annotation of home videos on mobile phones. It offers an MPEG-7-based home video database which can be queried in an intelligent and user-oriented manner exploiting users" personal domain ontologies. The platform employs terminal profiling techniques to deliver video retrieval user interfaces that personalize the search results according to the user's preferences and terminal capabilities, facilitating effective retrieval of home videos via various both mobile and fixed terminals. For video playout, Candela features a meta player, a video player augmented by an interactive metadata display which can be used for fast content-based in-video browsing, helping to avoid the consumption and streaming of uninteresting video parts, thus reducing network load. Thereby, Candela forms a comprehensive video management platform for mobile phones fully covering mobile home video management from acquisition to delivery.

This work studies a problem of personalization in the domain of content-based retrieval of home videos. The studies are made using a system prototype developed for content-based search of home videos, which allows recording, storing, annotating and showing of home videos on different platforms including mobile phones. This work presents results of a user study concerning the ways of shooting and showing home videos, identifies common scenarios and problems, and proposes simple ontology-based method for fast annotations and classification of videos recorded by e.g. mobile phones. Experimental results of video retrieval on the collected database of home videos show how this method facilitates real-time personalized home video presentations.

MediaTeam Oulu and VTT Technical Research Centre of Finland participated jointly in semantic feature extraction, manual search and interactive search tasks of TRECVID 2003. We participated to the semantic feature extraction by submitting results to 15 out of the 17 defined semantic categories. Our approach utilized spatio-temporal visual features based on correlations of quantized gradient edges and color values together with several physical features from the audio signal. Most recent version of our Video Browsing and Retrieval System (VIRE) contains an interactive cluster-temporal browser of video shots exploiting three semantic levels of similarity: visual, conceptual and lexical. The informativeness of the browser was enhanced by incorporating automatic speech transcription texts into the visual views based on shot key frames. The experimental results for interactive search task were obtained by conducting a user experiment of eight people with two system configurations: browsing by (I) visual features only (visual and conceptual browsing was allowed, no browsing with ASR text) or (II) visual features and ASR text (all semantic browsing levels were available and ASR-text content was visible). The interactive results using ASR-based features were better than the results using only visual features. This indicates the importance of successful integration of both visual and textual features for video browsing. In contrast to previous version of VIRE which performed early feature fusion by training unsupervised self-organizing maps, newest version capitalises on late fusion of features queries, which was evaluated in manual search task. This paper gives an overview of the developed system and summarises the results.