In this study, we have looked into the problem of vehicle detection in high-resolution satellite images. Based on the input from the local road authorities, we have focused not only on highways, but also on inner city roads, where more clutter is expected. The study site is the city of Oslo, Norway. To do vehicle detection in these areas, we propose an automatic approach, consisting of a segmentation step, followed by two stages of object classification. In the process, we utilize multispectral images, panchromatic images and a road network. The approach has been tested on Quickbird images, and the results that are obtained have been compared with manual counts and classifications.

Many documents contain (free-hand) underlining, "COPY" stamps, crossed out text, doodling and other "clutter" that occlude the text. In many cases, it is not possible to separate the text from the clutter. Commercial OCR solutions typically fail for cluttered text. We present a new method for finding the clutter using path analysis of points on the skeleton of the clutter/text connected component. This method can separate the clutter from the text even for fairly complex clutter shapes. Even with good localization of occluding clutter, it is difficult to use feature-based recognition for occluded characters, simply because the clutter affects the features in various ways. We propose a new algorithm that uses adapted templates of the font in the document that can be used for all forms of occlusion of the character. The method finds the simulated localization of the corresponding clutter in the templates and compares the unaffected parts of the templates and the character. The method has proved highly successful even when much of the character is occluded. We present examples of clutter localization and character recognition with occluded characters.

Perioperative mortality in coronary artery bypass grafting is usually caused by reduced left ventricular function due to regional myocardial ischemia or infarction. Post-operative graft occlusion is a well-known problem in coronary surgery. A sensitive tool to detect graft occlusion and monitor myocardial function may give the opportunity to revise malfunctioning grafts before departure from the hospital. This paper describes how a new method can detect cardiac ischemia using a 3-axis piezoelectric accelerometer. In three anesthetized pigs, a 3-axis piezoelectric accelerometer was sutured on the lateral free wall of the left ventricle. The left anterior descending (LAD) was occluded for different time periods and the accelerometer data were sampled with a PC. Short-time Fourier transform was calculated based on the accelerometer time series. The results were visualized using a 2D color-coded time-frequency plot. In the area of occlusion, a change to stronger power of higher harmonics was observed. Consequently, a difference value between the instant frequency pattern and a reference frequency pattern showed a rise in absolute value during the occlusion period. The preliminary results indicate that early recognition of regional cardiac ischemia is possible by analyzing accelerometer data acquired from the three animal trials using the prototype 3-axis accelerometer sensor.

The Norwegian company OptiNose AS is developing a novel concept that challenges traditional nasal delivery systems. The patented bi-directional delivery system takes advantage of the posterior connection between the nasal passages persisting when the soft palate automatically closes during oral exhalation. The exhalation into the delivery device triggers the release of particles into an airflow, which enters one nostril via a sealing nozzle and exits through the other nostril. During the development process for their drug delivery concept, OptiNose is using Computational Fluid Dynamics (CFD). The simulations are used to visualize and demonstrate the basic features of the bi-directional technique and discuss how its design and functionality could be further optimised. CFD computations thus increase the efficiency of device development and reduce the need for expensive and time consuming laboratory experiments and clinical tests. The simulations are carried out by use of Fluent, using a volume grid generated by Tgrid. To perform successful CFD calculations on the nose, construction of a proper surface grid of the nasal cavity, is mandatory. A high resolution CT-scan was the first part of this process, where after the 3-dimensional surface geometry was created by a pipeline of image segmentation and geometric modification steps. The process of building the surface grid is presented in the paper. Essential for the CFD computations is thorough validation of the simulation results. This paper presents comparisons between experimental and numerical simulations. Simple numerical models are used to verify critical parameters, e.g. airflow and particle size, in the design process, together with in vitro experiments in a human nasal airway model.

Ground cover classification based on a single satel- lite image can be challenging. The work reported here concerns the use of multitemporal satellite image data in order to alleviate this problem. We consider the problem of vegetation mapping and model the phenological evolution of the vegetation using a Hidden Markov Model (HMM). The different vegetation classes can be in one of a predefined set of states related to their phenological development. The characteristics of a given class are specified by the state transition probabilities as well as the probability of given satellite observations for that class and state. Classification of a specific pixel is thus reduced to selecting the class that has the highest probability of producing a given series of observations for that pixel. Compared to standard classification techniques such as maximum likelihood (ML) classification, the proposed scheme is flexible in that it derives its properties not only from image specific training data, but also from a model of the temporal behavior of the ground cover. It is shown to produce results that compare favorably to those obtained using ML classification on single satellite images, it also generalizes better than this approach. Obtaining good ground cover classifications based on a single satellite image can be challenging. The work reported here concerns the use of multitemporal satellite image data in order to alleviate this problem. We will consider an application of these methods to mapping of high mountain vegetation in Norway. The traditional mapping method based on manual field work is prohibitively expensive and alternatives are therefore sought. Vegetation classification based on satellite images is an interesting alternative, but the complexity of the vegetation ground cover is high and the use of multitemporal satellite image acquisitions is shown to improve the classifi- cation quality. This document is organized as follows: In the next section, we briefly recapitulate previous work related to multitemporal satellite image classification and phenological models. In section IV we discuss the HMM and how it can be used for classification. In section V we show results of the application of our algorithm, conclusions are given in section VI.

This work concerns the alignment of growth seasons based on satellite data. This work is motivated by a high mountain vegetation classification problem in Norway. Vegetation classes are characterized by their temporal evolution through a growth season. Data of high spatial resolution, like LANDSAT data, are often temporally sparse. In order to get a longer sequence of images, data from different years can be combined into one single synthetic sequence. We describe a method for determining the correspondence between the chronological time of the image acquisition and the time at which the phenological state of the vegetation cover shown in the image would typically occur. The task is considered as a minimization problem and is solved by dynamic programming. The methodology is based on the normalized difference vegetation index (NDVI) computed from data having a coarse spatial resolution such as MODIS or AVHRR data. The proposed methodology has been tested on data from several years covering a region in Norway including mountainous areas. It is evident from plots of the original data that NDVI curves from different seasons are shifted relative to one another. By applying the proposed time warping methodology to adjust the time scale within each year the shifts become less apparent. We conclude that the methodology can be used for alignment of growth seasons from satellite data.

Cryoablation is a method used for in situ destruction of liver tumors not eligible for surgical resection. Local recurrences following such treatment have been reported at rates of 5-44%. Insufficient procedural monitoring of the ablation is one plausible explanation for these recurrences. The cryoablative procedure is usually monitored by ultrasonography, but acoustic shadowing and loss of signals, compromise visualisation of the cryolesion circumference. Other monitoring modalities such as computer tomography and invasive methods like the use of thermocouples and impedance measurements have also been studied, but are not in common clinical use as single monitoring modalities. Thermodynamic conditions assumed adequate for tumor eradication are likely to occur only in parts of the cryolesion. This tumoricidal part of the cryolesion is not adequately depicted using any of these modalities. Magnetic resonance imaging (MRI) provides a clear delineation of the cryolesion circumference. Noninvasive temperature measurements assisted by MRI indicate which parts of the cryolesion that may be subject to complete necrosis. In this article MRI monitored cryoablation of liver tumors is discussed. Improved peroperative monitoring as offered by MRI may reduce the rates of local recurrences after treatment, but further technological improvements are required.

In this work, the aerodynamics and particle flow patterns in the human nose are simulated by Computational Fluid Dynamics (CFD). The simulations were carried out by use of Fluent, and the volume grid was generated by Tgrid. To be able to perform successful CFD-calculations of the nose, construction of a proper surface grid of the nasal cavity was mandatory. A high resolution CT-scan was the first part of this process, whereafter the 3-dimensional surface geometry was created by a pipeline of image segmentation andgeometric modification steps. The work described in this paper was initiated by a Norwegian company (OptiNose AS) developing a patented concept for efficient nasal delivery of drugs and vaccines. The objective of the article is to show how CFD can be applied as a tool to visualize and demonstrate the basic features of this technique and how it can be used to further optimize design and function of novel delivery devices based on the bi-directional concept. It is worth of notice, that the complexity of the nasal airway represents considerable challenges for CFD computations. This article shows, however, promising results on the issue. The simulations in Fluent are carried out with a reasonable range of flow rates and particle sizes typical to nasal delivery.

This study evaluate intraoperative Magnetic Resonance Imaging (MRI) as predictor of tissue damage following cryoablation of porcine liver with and without concomitant hepatic vascular inflow occlusion. Inflow occlusion was used during freezing in 6 of 12 pigs included. The volumes of the procedural ice-balls were estimated from MR images. Immediately after thawing contrast (MnDPDP) enhanced MRI was performed to estimate the volume of the cryolesion. Four days after ablation MRI was repeated of the in-vivo and the ex-vivo liver. Photography was performed of the sliced liver specimens to estimate the volumes of the lesions. The intraoperative volume of the cryolesion as shown by contrast enhanced MRI corresponded well to the ice-ball volume for lesions made without vascular occlusion (difference 0.3 +/- 0.9 cm(3), p = 0.239). For lesions made during occlusion the volume of the intraoperative cryolesion was larger than the corresponding ice-ball (difference 7.5 +/- 3.3 cm(3), p = 0.003). The volume of the cryolesions as estimated from histopathology four days after freezing and contrast enhanced MRI immediately after freezing corresponded well for lesions made with (difference -2.6 +/- 4.5 cm(3), p = 0.110) and without vascular occlusion (difference -0.5 +/- 2.3 cm(3), p = 0.695). Intraoperative MnDPDP-enhanced MRI of the cryolesion is predictive of the tissue damage induced during cryoablation of porcine liver. The procedural ice-ball is not, if induced during inflow occlusion.

The standard procedure for left anterior descending (LAD) coronary artery bypass requires a sternum split. The internal mammary artery (IMA), most typically the left one (LIMA), is dissected free from behind the sternum and is then anastomosed onto the heart below the occlusion of the LAD.Totally endoscopic coronary artery bypass (TECAB) is a less invasive alternative to the standard procedure. This procedure results in heavily reduced invasiveness, but also leads to loss of precision, reduced force feedback and loss of overview. The purpose of this work is to alleviate the problems related to loss of overview by generating an augmented reality for the surgeons in which they are given the impression of 'seeing through' the tissues surrounding the LIMA thus making localization of the LIMA a simple matter. Using preoperative CT or MR data, the LIMA is located with respect to the sternum and the tissues surrounding it. Intraoperatively, the sternum is located and the tissues surrounding the LIMA are tracked using a stereo videoscope held by a robotic arm. Knowing the position of the videoscope relative to the sternum now makes it possible to calculate where in the videoscopic images the LIMA is located. This information is then used to generate an augmented reality showing the tissues surrounding the LIMA as transparent, thus revealing the position of the LIMA within them. (C) 2003 Published by Elsevier Science B.V.

Chemical and isotope compositions of fluid samples, collected between 1974 and 1986 from 52 springs or shallow boreholes located in the Mont-Dore region (Massif Central, France), were examined. Some springs and wells were sampled several times during this period. The fluids emerge from Quaternary volcanic rocks or Paleozoic granite at temperatures between 4 and 62°C, and the origin of the H2O is meteoric. The waters can be classified into three groups: bicarbonate fluids, mixed bicarbonate-chloride fluids (with a mineralization up to 8 g/l), and acid-sulfate fluids. Only two fluids contain sufficient Cl− to be considered as ‘mature’ waters. Previous work has demonstrated that they all contain partly mantle-derived CO2 gas, and that the CO2-rich gas phase and bicarbonate-chloride waters are separated at substantial depth.Mineralized fluids circulate at depth and undergo several processes, such as cooling or dilution with recent freshwater, during their ascent to the surface. Therefore, the CO2-rich gas phase can be partly dissolved in the freshwater, or in deep fluids after their dilution. This process leads to the dissolution of surrounding rocks; such dissolution is discussed on the basis of major-element concentrations (Na, K, Ca, Mg), as well as the Sr 87/86 isotope ratio. Dissolution of S-bearing minerals has also been demonstrated. The presence of the CO2-rich gas phase also leads to isotope exchange between CO2 and H2O. Some mineralized fluids are less affected by these processes than others, in which case they display the chemical and isotopic characteristics of the original deep fluids.It was shown that the applicability of geothermometer calculations for these waters is hampered by several processes that modify the chemical composition. However, some geothermometers can be used for estimating the temperature of the deep fluids using the chemical composition of the less modified fluids. They indicate that fluids emerging from volcanic rocks in the Dordogne valley reach temperatures of around 100–130°C at depth, while the temperature of the fluid that issues from the granite at Saint-Nectaire is 160–175°C at depth.

Cryoablation is one of many local destruction techniques used to remove liver tumors. The procedure is cumbersome and time-consuming if there is no real-time access to intra-operative images as well as navigation information of instruments. In this paper, we propose a software-based solution that addresses issues like interactive, real-time navigation and visualization of cryoprobes, segmented tumor, intra-operative MR images, temperature maps, and isotherm surfaces. The system has an intuitive graphical user interface that provides access to a variety of functionalities. It enables the surgeon to perform the procedure safely, accurately, and rapidly by providing visual decision support mechanism.

Background. Local recurrences after cryoablation of liver tumors have been reported at rates from 5% to 44% and can be caused by inadequate coverage of the tumor by the frozen region. Hepatic vascular inflow occlusion may facilitate ablation by enlarging the size of the frozen region and the tissue necrosis induced by freezing. Few studies have documented these effects of inflow occlusion during liver cryoablation.Materials and methods. Two cryolesions were induced in the liver of 12 pigs in a standardized set-up. Vascular inflow occlusion was used in six pigs during freezing. Two freeze cycles were performed at each location. Ice-ball volume was estimated by intraoperative magnetic resonance imaging. Cryolesion volume was estimated from histopathologic examination of the lesions 4 days after ablation.Results. The median volume of ice-balls produced during inflow occlusion was 107% larger than for iceballs produced without occlusion (P < 0.001). The median volume of cryolesions made during inflow occlusion was 195% larger than for cryolesions induced without occlusion (P < 0.001). The geometry of the ice-balls was more regular if produced during inflow occlusion than if not. The ice-balls produced during the second freeze cycle were 17% and 20% larger than the ice-ball produced during the first freeze for lesions made with (P = 0.01) and without (P = 0.03) inflow occlusion.Conclusions. Hepatic vascular inflow occlusion enables freezing of larger volumes of liver tissue and increases the volume of tissue necrosis induced during cryoablation of porcine liver. (C) 2003 Elsevier Inc. All rights reserved.

This article reports on a method for planning probe placement in the liver during interventional procedures so as to avoid large hepatic vessels.

Continuous label segmentation approaches have recently attracted much interest as they provide a formalism for handling image artifacts due to the partial volume effect which is common in for instance medical images. Here, the authors propose a new approach to this type of segmentation. Their work represents an extension of the now classic Markovian Bayesian discrete label segmentation approaches and provides good results on synthetic images simulating the presence of partial volumes as well as on real patient MR images