Anders Bjorholm Dahl

Assessment of algorithms for mitosis detection in breast cancer histopathology images.

The proliferative activity of breast tumors, which is routinely estimated by counting of mitotic figures in hematoxylin and eosin stained histology sections, is considered to be one of the most important prognostic markers. However, mitosis counting is laborious, subjective and may suffer from low inter-observer agreement. With the wider acceptance of whole slide images in pathology labs, automatic image analysis has been proposed as a potential solution for these issues. In this paper, the results from the Assessment of Mitosis Detection Algorithms 2013 (AMIDA13) challenge are described. The challenge was based on a data set consisting of 12 training and 11 testing subjects, with more than one thousand annotated mitotic figures by multiple observers. Short descriptions and results from the evaluation of eleven methods are presented. The top performing method has an error rate that is comparable to the inter-observer agreement among pathologists.

Genetic analyses of the human eye colours using a novel objective method for eye colour classification

In this study, we present a new objective method for measuring the eye colour on a continuous scale that allows researchers to associate genetic markers with different shades of eye colour. With the use of the custom designed software Digital Iris Analysis Tool (DIAT), the iris was automatically identified and extracted from high resolution digital images. DIAT was made user friendly with a graphical user interface. The software counted the number of blue and brown pixels in the iris image and calculated a Pixel Index of the Eye (PIE-score) that described the eye colour quantitatively. The PIE-score ranged from -1 to 1 (brown to blue). The software eliminated the need for user based interpretation and qualitative eye colour categories. In 94% (570) of 605 analyzed eye images, the iris region was successfully extracted and a PIE-score was calculated. A very high correlation between the PIE-score and the human perception of eye colour was observed. The correlations between the PIE-scores and the six IrisPlex SNPs (HERC2 rs12913832, OCA2 rs1800407, SLC24A4 rs12896399, TYR rs1393350, SLC45A2 rs16891982 and IRF4 rs12203592) were analyzed in 570 individuals. Significant differences (p<10(-6)) in the PIE-scores of the individuals typed as HERC2 rs12913832 G (PIE=0.99) and rs12913832 GA (PIE=-0.71) or A (PIE=-0.87) were observed. We adjusted for the effect of HERC2 rs12913832 and showed that the quantitative PIE-scores were significantly associated with SNPs with minor effects (OCA2 rs1800407, SLC24A4 rs12896399 and TYR rs1393350) on the eye colour. We evaluated the two published prediction models for eye colour (IrisPlex [1] and Snipper[2]) and compared the predictions with the PIE-scores. We found good concordance with the prediction from individuals typed as HERC2 rs12913832 G. However, both methods had difficulties in categorizing individuals typed as HERC2 rs12913832 GA because of the large variation in eye colour in HERC2 rs12913832 GA individuals. With the use of the DIAT software and the PIE-score, it will be possible to automatically compare the iris colour of large numbers of iris images obtained by different studies and to perform large meta-studies that may reveal loci with small effects on the eye colour.

Spectral characterisation of dairy products using photon time-of-flight spectroscopy

In this paper, we present, for the first time, the absorption and reduced scattering spectra of commercially available milk and yoghurt products, obtained using photon-time-of-flight spectroscopy. The ability of this technique to separate the contributions from absorption and scattering in the sample provides important information on the chemical composition and micro-structural properties, which are not available with the traditional techniques used in dairy production. The instrument operates in the spectral range from 500 nm to 1030 nm. The reduced scattering coefficient varies from 5 cm−1 for milk with 0.1% fat in the near infrared range, to 60 cm−1 for yoghurt with 3.0% fat in the green wavelength regime. The absorption is within the range of 0.05–0.5cm−1, with only small variation in the absolute value between products. Our results show that the reduced scattering clearly distinguishes milk and yoghurt with the same fat content and can offer a reliable way of monitoring structural formation during milk fermentation.

Large-Scale Data for Multiple-View Stereopsis

The seminal multiple-view stereo benchmark evaluations from Middlebury and by Strecha et al. have played a major role in propelling the development of multi-view stereopsis (MVS) methodology. The somewhat small size and variability of these data sets, however, limit their scope and the conclusions that can be derived from them. To facilitate further development within MVS, we here present a new and varied data set consisting of 80 scenes, seen from 49 or 64 accurate camera positions. This is accompanied by accurate structured light scans for reference and evaluation. In addition all images are taken under seven different lighting conditions. As a benchmark and to validate the use of our data set for obtaining reasonable and statistically significant findings about MVS, we have applied the three state-of-the-art MVS algorithms by Campbell et al., Furukawa et al., and Tola et al. to the data set. To do this we have extended the evaluation protocol from the Middlebury evaluation, necessitated by the more complex geometry of some of our scenes. The data set and accompanying evaluation framework are made freely available online. Based on this evaluation, we are able to observe several characteristics of state-of-the-art MVS, e.g. that there is a tradeoff between the quality of the reconstructed 3D points (accuracy) and how much of an object’s surface is captured (completeness). Also, several issues that we hypothesized would challenge MVS, such as specularities and changing lighting conditions did not pose serious problems. Our study finds that the two most pressing issues for MVS are lack of texture and meshing (forming 3D points into closed triangulated surfaces).

Comparison of a multispectral vision system and a colorimeter for the assessment of meat color

The color assessment ability of a multispectral vision system is investigated by a comparison study with color measurements from a traditional colorimeter. The experiment involves fresh and processed meat samples. Meat is a complex material; heterogeneous with varying scattering and reflectance properties, so several factors can influence the instrumental assessment of meat color. In order to assess whether two methods are equivalent, the variation due to these factors must be taken into account. A statistical analysis was conducted and showed that on a calibration sheet the two instruments are equally capable of measuring color. Moreover the vision system provides a more color rich assessment of fresh meat samples with a glossier surface, than the colorimeter. Careful studies of the different sources of variation enable an assessment of the order of magnitude of the variability between methods accounting for other sources of variation leading to the conclusion that color assessment using a multispectral vision system is superior to traditional colorimeter assessments.

Enterococcus faecalis infective endocarditis: focus on clinical aspects

Enterococcus faecalis infective endocarditis (IE) is a disease of increasing importance, with more patients infected, increasing frequency of health-care associated infections and increasing incidence of antimicrobial resistances. The typical clinical presentation is a subacute course with fever, malaise and generalized aches, difficult to distinguish from other more common diseases. Of paramount importance is transthoracic- and transesophageal-echocardiography to establish the diagnosis. At the moment, the predominant strategies recommend ampicillin in combination with either gentamicin or ceftriaxone. E. faecalis infective endocarditis continues to be a very serious disease with considerable percentages of high-level gentamicin resistant strains and in-hospital mortality around 20%. Strategies to prevent E. faecalis IE, improve diagnostics, optimize treatment and reduce morbidity will be necessary to improve the overall prognosis.

Non-invasive assessment of dairy products using spatially resolved diffuse reflectance spectroscopy.

The quality of a dairy product is largely determined by its microstructure which also affects its optical properties. Consequently, an assessment of the optical properties during production may be part of a feedback system for ensuring the quality of the production process. This paper presents a novel camera-based measurement technique that enables robust quantification of a wide range of reduced scattering coefficients and absorption coefficients. Measurements are based on hyperspectral images of diffuse reflectance in the wavelength range of 470 to 1020 nm. The optical properties of commercially available milk and yogurt products with three different levels of fat content are measured. These constitute a relevant range of products at a dairy plant. The measured reduced scattering properties of the samples are presented and show a clear discrimination between levels of fat contents as well as fermentation. The presented measurement technique and method of analysis is thus suitable for a rapid, non-contact, and non-invasive inspection that can deduce physically interpretable properties.

Prediction of pork quality parameters by applying fractals and data mining on MRI

This work firstly investigates the use of MRI, fractal algorithms and data mining techniques to determine pork quality parameters non-destructively. The main objective was to evaluate the capability of fractal algorithms (Classical Fractal algorithm, CFA; Fractal Texture Algorithm, FTA and One Point Fractal Texture Algorithm, OPFTA) to analyse MRI in order to predict quality parameters of loin. In addition, the effect of the sequence acquisition of MRI (Gradient echo, GE; Spin echo, SE and Turbo 3D, T3D) and the predictive technique of data mining (Isotonic regression, IR and Multiple linear regression, MLR) were analysed. Both fractal algorithm, FTA and OPFTA are appropriate to analyse MRI of loins. The sequence acquisition, the fractal algorithm and the data mining technique seems to influence on the prediction results. For most physico-chemical parameters, prediction equations with moderate to excellent correlation coefficients were achieved by using the following combinations of acquisition sequences of MRI, fractal algorithms and data mining techniques: SE-FTA-MLR, SE-OPFTA-IR, GE-OPFTA-MLR, SE-OPFTA-MLR, with the last one offering the best prediction results. Thus, SE-OPFTA-MLR could be proposed as an alternative technique to determine physico-chemical traits of fresh and dry-cured loins in a non-destructive way with high accuracy.

Noninvasive particle sizing using camera-based diffuse reflectance spectroscopy

Diffuse reflectance measurements are useful for noninvasive inspection of optical properties such as reduced scattering and absorption coefficients. Spectroscopic analysis of these optical properties can be used for particle sizing. Systems based on optical fiber probes are commonly employed, but their low spatial resolution limits their validity ranges for the coefficients. To cover a wider range of coefficients, we use camera-based spectroscopic oblique incidence reflectometry. We develop a noninvasive technique for acquisition of apparent particle size distributions based on this approach. Our technique is validated using stable oil-in-water emulsions with a wide range of known particle size distributions. We also measure the apparent particle size distributions of complex dairy products. These results show that our tool, in contrast to those based on fiber probes, can deal with a range of optical properties wide enough to track apparent particle size distributions in a typical industrial process.

Simultaneous tomographic reconstruction and segmentation with class priors

We consider tomographic imaging problems where the goal is to obtain both a reconstructed image and a corresponding segmentation. A classical approach is to first reconstruct and then segment the image; more recent approaches use a discrete tomography approach where reconstruction and segmentation are combined to produce a reconstruction that is identical to the segmentation. We consider instead a hybrid approach that simultaneously produces both a reconstructed image and segmentation. We incorporate priors about the desired classes of the segmentation through a Hidden Markov Measure Field Model, and we impose a regularization term for the spatial variation of the classes across neighbouring pixels. We also present an efficient implementation of our algorithm based on state-of-the-art numerical optimization algorithms. Simulation experiments with artificial and real data demonstrate that our combined approach can produce better results than the classical two-step approach.