UmBio AB

Umeå, Sweden
Umeå, Sweden
Time filter
Source Type

Vermaak I.,Tshwane University of Technology | Viljoen A.,Tshwane University of Technology | Lindstrom S.W.,Swedish Defence Research Agency | Lindstrom S.W.,UmBio AB
Journal of Pharmaceutical and Biomedical Analysis | Year: 2013

Illicium verum (Chinese star anise) dried fruit is popularly used as a remedy to treat infant colic. However, instances of life-threatening adverse events in infants have been recorded after use, in some cases due to substitution and/or adulteration of I. verum with Illicium anisatum (Japanese star anise), which is toxic. It is evident that rapid and efficient quality control methods are of utmost importance to prevent re-occurrence of such dire consequences. The potential of short wave infrared (SWIR) hyperspectral imaging and image analysis as a rapid quality control method to distinguish between I. anisatum and I. verum whole dried fruit was investigated. Images were acquired using a sisuChema SWIR hyperspectral pushbroom imaging system with a spectral range of 920-2514nm. Principal component analysis (PCA) was applied to the images to reduce the high dimensionality of the data, remove unwanted background and to visualise the data. A classification model with 4 principal components and an R2X_cum of 0.84 and R2Y_cum of 0.81 was developed for the 2 species using partial least squares discriminant analysis (PLS-DA). The model was subsequently used to accurately predict the identity of I. anisatum (98.42%) and I. verum (97.85%) introduced into the model as an external dataset. The results show that SWIR hyperspectral imaging is an objective and non-destructive quality control method that can be successfully used to identify whole dried fruit of I. anisatum and I. verum. In addition, this method has the potential to detect I. anisatum whole dried fruits within large batches of I. verum through upscaling to a conveyor belt system. © 2012 Elsevier B.V.

Lindstrom S.W.,Umeå University | Lindstrom S.W.,UmBio AB | Geladi P.,Swedish University of Agricultural Sciences | Jonssonb O.,UmBio AB | Petterssonb F.,UmBio AB
Journal of Near Infrared Spectroscopy | Year: 2011

This study investigates the effect of imbalanced spectral data in the training set, when developing partial least squares discriminant analysis (PLS-DA) classification models for use in future predictions. The experimental study was performed using a real hyperspectral short-wavelength infrared image data set collected from bakery products (buns) containing contaminants (flies) but similar applications for other insects, paper and plastic were also tested. The contaminants represent a very small proportion of the images relative to the bun. The PLS-DA model aims at accurately detecting and classifying the contaminants and this requires a modification of the calibration data set. The paper deals with problems caused by unbalanced calibration data sets and how to remedy them. In the example it was demonstrated that, by balancing the calibration data from 58,476 bun pixels + 279 fly pixels to 279 bun + 279 fly pixels, the number of true predictions could be improved with a smaller number of PLS components used in the model. The improvement for flies increased from 65% true predictions with ten PLS components to 99% true prediction with five to six PLS components. The true prediction for bun went from 100% to 99.5% with six PLS components which is an acceptable reduction. Theoretical explanations are included. © 2011 IM Publications LLP. All rights reserved.

Boden I.,Umeå University | Larsson W.,Umeå University | Nilsson D.,UmBio AB | Forssell E.,Umeå University | And 2 more authors.
Skin Research and Technology | Year: 2011

Background/purpose: Near-infrared (NIR) spectroscopy and skin impedance (IMP) measurements are useful techniques for objective diagnostics of various skin diseases. Here, we present a combined probe head for simultaneous, time-saving NIR spectroscopy and skin impedance measurements. The probe also ensures that both measurements are performed under equal conditions and at the same skin location. Methods: Finite element method simulations were performed for evaluation of the impedance. In vivo skin measurements were performed and combined NIR and impedance spectra were analysed by means of multivariate methods with respect to body location, age and gender. The classification rate was determined by a planar discriminant analysis. Reproducibility was investigated by calculation of scatter values and statistical significance between overlapping groups was assessed by the calculation of intra-model distances, q. Results: The novel probe yielded rapid reproducible results and was easy to manage. Significant differences between skin locations and to a lesser extent age groups and gender were demonstrated. Conclusion: With the novel probe, statistically significant differences between overlapping classes in score plots can be confirmed by calculating intra-model distances. The influence of molecular differences in the skin at different body locations is larger than the influence of gender or age and therefore relevant reference measurements are discussed. © 2011 John Wiley & Sons A/S.

A method for determining an estimate of a grouping of a plurality of objects (104) into at least one group according to extent of relatedness of the objects is disclosed. A set of data of each of the plurality of objects is acquired or determined, wherein the set of data of each of the objects is based on spectral information of the object. Based on the set of data of each object, an initial estimate of the grouping of the plurality of objects is determined. Starting from the initial estimate of the grouping of the plurality of objects, at least one further estimate of the grouping of the plurality of objects is iteratively determined, until a difference between values of at least one fitness function based on an estimate of the grouping of the plurality of objects and a preceding estimate of the grouping of the plurality of objects complies with a selected criterion for change in the value of the at least one fitness function. The iterative determination of at least one further estimate of the grouping of the plurality of objects comprises, for a preceding estimate of the grouping of the plurality of objects, determining a statistical relation between the set of data of each of the objects and information relating to in which group each of the objects is included, determining a measure of goodness of the statistical relation, wherein the value of the at least one fitness function is based on the determined measure, and determining a new estimate of the grouping of the plurality of objects based on the value of the at least one fitness function.

Loading UmBio AB collaborators
Loading UmBio AB collaborators