Quality Control for Scalable Media Processing Applications

Many media processing applications create a load that varies significantly over time. Hence, if such an application is assigned a lower processing-time budget than needed in its worst-case load situation, deadline misses are likely to occur. This problem can be dealt with by designing media processing applications in a scalable fashion. A scalable media processing application can run in multiple qualities, leading to correspondingly different resource demands. The problem we consider is to find an accompanying quality control strategy, which minimizes both the number of deadline misses and the number of quality changes, while maximizing the quality of processing. We present an initial approach to the above problem by modeling it as a Markov decision process (MDP). Our model is based on measuring relative progress at milestones. Solving the MDP results in a quality control strategy that can be applied during runtime with only little overhead. We evaluate our approach by means of a practical example, which concerns a scalable MPEG-2 decoder.     

A comparison of algorithms for control constrained optimal control of the Burgers equation

Abstract A comparison of three different but related numerical methods for control constrained optimal control of the Burgers equation is carried out. We develop the principal ideas of the different strategies considered and present detailed numerical examples. Lastly, conclusions about the behavior of the approaches are obtained.     

Impact of dimension uncertainty and model calibration on sewer system assessment.

Assessments of sewer performance are usually based on a single computation of CSO (combined sewer overflow) volumes using a time series of rainfall as system loads. A shortcoming of this method is that uncertainties in knowledge of sewer system dimensions are not taken into account. Moreover, sewer models are rarely calibrated. This paper presents the impacts of database errors and model calibration on return periods of calculated CSO volumes. The impact of uncertainties is illustrated with two examples. Variability of calculated CSO volumes is estimated using Monte Carlo simulations. The results show that calculated CSO volumes vary considerably due to database errors, especially uncertain dimensions of the catchment area. Furthermore, event-based calibration of a sewer model does not result in more reliable predictions because the calibrated parameters have low portability. However, it enables removal of database errors harmonising model predictions and 'reality'.     

Possibilities and limits of extracranial vascular diagnosis with conventional echography

OBJECTIVE: To evaluate the efficacy of terazosin, an alpha-blocker, for the treatment of idiopathic oligozoospermia. PATIENTS AND METHODS: Thirty couples with infertility whose only detectable abnormality was male idiopathic subfertility entered the study. The diagnosis of idiopathic subfertility in all males studied, aged 26 to 38 years (mean 28.2 years), was confirmed after exclusion of any iatrogenic, systemic, congenital, infectious, autoimmune or endocrinological cause. In order to start with a baseline value before the study, at least three semen samples were evaluated in accordance with the WHO recommendation. Before initiation of treatment, blood samples were drawn for measurement of FSH, LH, testosterone, prolactin, dihydrotestosterone, and estradiol. Fifteen randomly selected patients (Group A) received 2 mg/d of alpha-blocker (terazosin), while another 15 (Group B) were administered an identically packed placebo tablet. Both groups received therapy for 6 months. RESULTS: The mean seminal volume changed insignificantly between the two groups (4.15 +/- 1.95 vs. 4.10 +/- 1.95). There was a statistically significant increase of the sperm concentration in patients who received the alpha-blocker compared to those receiving placebo (24.76 +/- 9.45 vs. 13.15 +/- 11.55 millions/mL; P < .001). No improvement of the mean percentage of abnormal spermatozoa was observed in the treated patients, nor a statistically significant difference of sperm motility in the treated group compared to the placebo group. Side effects were not observed in the patients receiving terazosin treatment, or were so minimal that therapy was continued. The pregnancy rates did not differ between the two groups to a statistically significant degree. CONCLUSION: The administration of terazosin to patients with idiopathic oligozoospermia has a demonstrably positive effect, especially on sperm concentration.     

Increased expression of peripheral benzodiazepine receptors in the facial nucleus following motor neuron axotomy

Peripheral benzodiazepine receptors (PBRs) are expressed in a variety of tissues but are normally found at low levels in the brain. Following various types of nerve injury, a reactive gliosis results that exhibits a high expression of this receptor. To further characterize the expression of PBRs following neuronal injury, we evaluated PBR expression in the facial nucleus following facial nerve axotomy (FNA). Injury to a peripheral nerve results in a complex series of metabolic and morphological changes around the injured neuron. Transections of the facial nerve results in a rapid activation of both astrocytes and microglia around axotomized motor neurons. FNA resulted in an increase in the staining for both astrocytes (glial fibrillary acidic protein) and activated microglia (OX42). There was also a reduction in synaptic contacts with the motor nucleus as evidenced by reduced staining for the synaptic marker, synaptophysin. In sections labeled with [3H]-PK11195, the subsequent autoradiograms displayed marked increases in the labeling for PBRs. This increase was observed at 5, 7 and 10 days after nerve transection. The increase was primarily in the level of expression (Bmax), with no change in the affinity of the ligand (Kd). The increase in PBR expression after FNA supports the hypothesis that PBRs can be used as a sensitive marker for CNS injury.     

A tool to evaluate the fertiliser value and the environmental impact of substrates from wastewater treatment.

Fertilisers may contain pollutants that are applied to the field together with the nutrients. Comparing fertilisers is difficult because of their different concentrations of nutrients and pollutants. In this study an already existing model was taken. It was further developed to compare nutrient fluxes (N, P, K, Ca, Mg, S, humus) and pollutants (heavy metals and pharmaceuticals) of pig and cattle slurry as well as human urine. The data used is taken from literature and, in the case of pharmaceuticals in urine, daily excretion rates were calculated. An amount of 19 m3 ha(-1) a(-1) undiluted urine can be applied, limiting factor is sulphur. Without taking over-fertilisation into account, an addition of mineral fertiliser is required to any organic fertiliser application. In general, heavy metal, antibiotic, and hormone fluxes are higher by using animal manure then urine as fertiliser. However, additional loads of other pharmaceuticals consumed by humans have to be considered. Overall, the model is a suitable tool worthwhile to be extended in aspects of hygiene, environmental impacts as for example, degradation processes as well as ecotoxicology.     

Space and time bounds on indexing 3D models from 2D images

Model-based visual recognition systems often match groups of image features to groups of model features to form initial hypotheses, which are then verified. In order to accelerate recognition considerably, the model groups can be arranged in an index space (hashed) offline such that feasible matches are found by indexing into this space. For the case of 2D images and 3D models consisting of point features, bounds on the space required for indexing and on the speedup that such indexing can achieve are demonstrated. It is proved that, even in the absence of image error, each model must be represented by a 2D surface in the index space. This places an unexpected lower bound on the space required to implement indexing and proves that no quantity is invariant for all projections of a model into the image. Theoretical bounds on the speedup achieved by indexing in the presence of image error are also determined, and an implementation of indexing for measuring this speedup empirically is presented. It is found that indexing can produce only a minimal speedup on its own. However, when accompanied by a grouping operation, indexing can provide significant speedups that grow exponentially with the number of features in the groups