Deep learning for content-based video retrieval in film and television production

While digitization has changed the workflow of professional media production, the content-based labeling of image sequences and video footage, necessary for all subsequent stages of film and television production, archival or marketing is typically still performed manually and thus quite time-consuming. In this paper, we present deep learning approaches to support professional media production. In particular, novel algorithms for visual concept detection, similarity search, face detection, face recognition and face clustering are combined in a multimedia tool for effective video inspection and retrieval. The analysis algorithms for concept detection and similarity search are combined in a multi-task learning approach to share network weights, saving almost half of the computation time. Furthermore, a new visual concept lexicon tailored to fast video retrieval for media production and novel visualization components are introduced. Experimental results show the quality of the proposed approaches. For example, concept detection achieves a mean average precision of approximately 90% on the top-100 video shots, and face recognition clearly outperforms the baseline on the public Movie Trailers Face Dataset.     

Secure mobile communication via identity-based cryptography and server-aided computations

In this paper, an identity-based key agreement protocol for securing mobile telephony in GSM and UMTS networks is presented. The approach allows two mobile phones to perform a session key agreement over an unsecured channel and between different providers using telephone numbers as public keys. Using the created session key, a symmetric encryption of all call data can be performed. Solutions to the problems of multi-domain key generation, key distribution, multi-domain public parameter distribution and inter-domain key agreement are presented. Furthermore, the proposed approach can be speeded up using server-aided cryptography, by outsourcing computationally expensive cryptographic operations to a high-performance backend computing server. The feasibility of the approach is illustrated by presenting experimental results based on a Symbian implementation running on N95-1 and N82-1 Nokia smartphones.     

Fitness landscape analysis and memetic algorithms for the quadratic assignment problem

In this paper, a fitness landscape analysis for several instances of the quadratic assignment problem (QAP) is performed, and the results are used to classify problem instances according to their hardness for local search heuristics and meta-heuristics based on local search. The local properties of the fitness landscape are studied by performing an autocorrelation analysis, while the global structure is investigated by employing a fitness distance correlation analysis. It is shown that epistasis, as expressed by the dominance of the flow and distance matrices of a QAP instance, the landscape ruggedness in terms of the correlation length of a landscape, and the correlation between fitness and distance of local optima in the landscape together are useful for predicting the performance of memetic algorithms-evolutionary algorithms incorporating local search (to a certain extent). Thus, based on these properties, a favorable choice of recombination and/or mutation operators can be found. Experiments comparing three different evolutionary operators for a memetic algorithm are presented.     

Fitness landscapes, memetic algorithms, and greedy operators for graph bipartitioning

The fitness landscape of the graph bipartitioning problem is investigated by performing a search space analysis for several types of graphs. The analysis shows that the structure of the search space is significantly different for the types of instances studied. Moreover, with increasing epistasis, the amount of gene interactions in the representation of a solution in an evolutionary algorithm, the number of local minima for one type of instance decreases and, thus, the search becomes easier. We suggest that other characteristics besides high epistasis might have greater influence on the hardness of a problem. To understand these characteristics, the notion of a dependency graph describing gene interactions is introduced. In particular, the local structure and the regularity of the dependency graph seems to be important for the performance of an algorithm, and in fact, algorithms that exploit these properties perform significantly better than others which do not. It will be shown that a simple hybrid multi-start local search exploiting locality in the structure of the graphs is able to find optimum or near optimum solutions very quickly. However, if the problem size increases or the graphs become unstructured, a memetic algorithm (a genetic algorithm incorporating local search) is shown to be much more effective.     

Ursodeoxycholate stabilizes phospholipid-rich membranes and mimics the effect of cholesterol: investigations on large unilamellar vesicles

Ichthyophis kohtaoensis, a member of the limbless Gymnophiona, has a specialized subterranean burrowing mode of life and a predominantly olfactory-guided orientation. The only visually guided behavior seems to be negative phototaxis. As these animals possess extremely small eyes (only 540 microm in diameter in adults), functional investigations of single retinal cells by electrophysiological methods have so far failed. Therefore, the content and distribution of retinal transmitters have been investigated as indications for a functioning sense organ in an animal that is supposed to be blind. In this study, the organization and development of the dopaminergic system have been examined in the retinae of embryonic, larval, and adult I. kohtaoensis, by using an antiserum against tyrosine hydroxylase, the rate-limiting enzyme in the catecholamine synthetic pathway, and an antiserum against dopamine itself. Labeled somata are situated in the inner nuclear layer and in the ganglion cell layer. Dopamine-positive fibers form a dense diffuse plexus, that covers the whole inner plexiform layer, whereas tyrosine hydroxylase-immunoreactive processes show a tendency to arborize in a stratified manner. Tyrosine-hydroxylase-immunolabeled fibers can occasionally be observed in the optic nerve head of larval stages. During ontogenesis and larval development, the distribution of transmitter-expressing cells changes and their number decreases, but no general degeneration of the visual system is detectable. Adult Ichthyophis still have retinal transmitters, indicating that the eyes, although obviously playing a minor role in a subterranean ecological niche, retain all the elements of functioning sense organs.     

Time-constrained services: a framework for using real-time web services in industrial automation

The use of web services in industrial automation, e.g. in fully automated production processes like car manufacturing, promises simplified interaction among the manufacturing devices due to standardized protocols and increased flexibility with respect to process implementation and reengineering. Moreover, the adoption of web services as a seamless communication backbone within the overall industrial enterprise has additional benefits, such as simplified interaction with suppliers and customers (i.e. horizontal integration) and avoidance of a break in the communication paradigm within the enterprise (i.e. vertical integration). The Time-Constrained Services (TiCS) framework is a development and execution environment that empowers automation engineers to develop, deploy, publish, compose, and invoke time-constrained web services. TiCS consists of four functional layers—tool support layer, real-time infrastructural layer, real-time service layer, and hardware layer—which contain several components to meet the demands of a web service based automation infrastructure. This article gives an overview of the TiCS framework. More precisely, the general design considerations and an architectural blueprint of the TiCS framework are presented. Subsequently, selected key components of the TiCS framework are discussed in detail: the SOAP4PLC engine for equipping programmable logic controllers with a web service interface, the SOAP4IPC engine for processing web services in real-time on industrial PCs, the WS-TemporalPolicy language for describing time constraints, and the TiCS Modeler for composing time-constrained web services into a time-constrained BPEL4WS workflow.     

PD source identification with novel discharge parameters usingcounterpropagation neural networks

Computer aided partial discharge (PD) source identification using different multidimensional discharge patterns is widely regarded as an important tool for insulation diagnosis. In this paper, a neural network (NN) approach to PD pattern classification is presented. The approach is based on applying variants of the counterpropagation NN architecture to the classification of PD patterns. These patterns are derived from physically related discharge parameters, different from those commonly used. It is shown that considerable improvements of the classification quality can be obtained when an extended counterpropagation network with a dynamically changing network topology is applied to patterns that employ the voltage difference between consecutive pulses instead of the phase of occurrence as the main discharge parameter. Furthermore, using a particular parameter vector that takes the correlation between consecutive discharges into account also allows to solve the rejection problem with this type of NN     

Using counterpropagation neural networks for partial discharge diagnosis

In high voltage engineering, various methods of non-destructive fault diagnosis are applied for investigating the quality of insulating materials and systems. The methods are aimed at classifying patterns derived from the measured characteristics of the electrical signals typically resulting from insulation defects. In this paper, variants of the counterpropagation neural network architecture are used to classify patterns representing various properties of partial discharges. It is shown that the classification quality can be improved considerably when an extended counterpropagation network with a dynamically changing network topology, and an additional vigilance unit for monitoring the behaviour of the network during the learning phase is applied. The extended network has significant advantages over the standard counterpropagation network in cases where outliers in the training data seriously degrade the approximation quality of the standard network. When using the proposed network in conjunction with physically motivated discharge data, input patterns from defect categories not considered during training can be rejected more reliably. This rejection problem is particularly important for practical applications where misclassifications cannot be tolerated.     

Interaction of glucose and metformin with isolated red cell membrane

Isolated human erythrocyte membranes (red blood cell (RBC) ghosts) were incubated with glucose at 5, 10, 20 and 100 mmol/l concentrations, with insulin (0.01 to 200 mU/l) and metformin (CAS 657-24-9) 0.5 up to 50.0 mumol/l). Binding studies with 14C-glucose and subsequent gel electrophoresis revealed 60% of the radioactivity around ban 4.2-4.5 at 5 mmol/l, whereas a random distribution of radioactivity over all protein bands of the RBC membrane was found at 20 mmol/l concentration after incubation for 30 min or 48 h. Metformin does not bind covalently to RBC membranes, however, after photochemical linkage of 14C-metformin via the aminoreactive linker azidophenylglyoxal the highest radioactivity (21%) was counted in the range of band 4.2-4.5. In parallel with an increase of order parameters of 5-doxyl-stearic acid the thiol status of the membranes decreases as determined by monobromobimane fluorescence. 20 and 100 mmol/l concentrations of glucose decrease the reactivity of membrane thiols towards bromobimane significantly to 73 and 62% of the controls. Concomitantly, membrane fluidity at polar sites is diminished as measured by order parameters of spin label 5-doxyl stearic acid. In RBC membranes pretreated with 20 mmol/l glucose the decreased fluorescence is significantly raised again by insulin and metformin. This effect is even more pronounced, if insulin and metformin are incubated together. Reaction of membrane thiols with a maleimido spin label detects modification in the ratio of mobile and immobilized spin label populations in the electron paramagnetic resonance signal under the above conditions, indicative of conformational changes of membrane proteins.