Koninklijke Philips Corporation N.v.

Bart Jacob Bakker

Daniel Gagnon

Frederik Jan De Bruijn

Gerhardus Wilhelmus Lucassen

Harold Agnes Wilhelmus Schmeitz

Dimitrios Mavroeidis

Rene Van Den Ham

Hendrik Jan Van Ooijen

Stojan Trajanovski

Lothar Spies

Manoj Narayanan

Alexander Fischer

Amir Hussein Rmaile

Marinus Karel Johannes De Jager

Ruud Johannes Gerardus Van Sloun

John Taylor

Philip Preshaw

Jenny Margarito

René Van Den Ham

Sergey Kurkin

Henk Jan Van Ooijen

Jinhui Qian

Vincent Maurice André Auvray

Ren Van Den Ham

Dimitros Mavrieudus

The invention relates a computer-implemented method () of generating explainability information for explaining a model output of a trained model. The method uses one or more aspect recognition models configured to indicate a presence of respective characteristics in the input instance. A saliency method is applied to obtain a masked source representation of the input instance at a source layer of the trained model (e.g., the input layer or an internal layer), comprising those elements at the source layer relevant to the model output. The masked source representation is mapped to a target layer (e.g., input or internal layer) of an aspect recognition model, and the aspect recognition model is then applied to obtain a model output indicating a presence of the given characteristic relevant to the model output of the trained model. As explainability information, the characteristics indicated by the aspect recognition models are output.

Explaining a model output of a trained model

The invention relates to a method for predicting a value of a parameter of a system (). The value of the parameter of the system () is predicted by incrementally optimizing a pharmacokinetic and pharmacodynamic model based on values of parameters of the system () received over time. This allows for predicting values of the parameters of the system () with improved accuracy. In one embodiment the system () comprises a coagulation system () comprising an anticoagulant. The predicted value of the parameter of the system () can be a point in time at which the coagulation system () reaches hemostatic balance after a periodic supply of anticoagulant to the coagulation system () is discontinued.

Incrementally optimized pharmacokinetic and pharmacodynamic model

The present invention provides a method, computer program and processing system for identifies boundaries of lesions within image data. The image data is processed using a machine learning algorithm to generate probability data and uncertainty data. The probability data provides, for each image data point of the image data, a probability data points indicating a probability that said image data point is part of a lesion. The uncertainty data provides, for each probability data point, an uncertainty data point indicating an uncertainty of the said probability data point. The uncertainty data is processed to identify or correct boundaries of the lesions.

Identifying boundaries of lesions within image data

Aspects and embodiments relate to a method of providing a representation of a feature identified by a deep neural network as being relevant to an outcome, a computer program product and apparatus configured to perform that method. The method comprises: providing the deep neural network with a training library comprising: a plurality of samples associated with the outcome; using the deep neural network to recognise a feature in the plurality of samples associated with the outcome; creating a feature recognition library from an input library by identifying one or more elements in each of a plurality of samples in the input library which trigger recognition of the feature by the deep neural network; using the feature recognition library to synthesise a plurality of one or more elements of a sample which have characteristics which trigger recognition of the feature by the deep neural network; and using the synthesised plurality of one or more elements to provide a representation of the feature identified by the deep neural network in the plurality of samples associated with the outcome. Accordingly, rather than visualising a single instance of one or more elements in a sample which trigger a feature associated with an outcome, it is possible to visualise a range of samples including elements which would trigger a feature associated with an outcome, thus enabling a more comprehensive view of operation of a deep neural network in relation to a particular feature.

Deep neural network visualisation

Disclosed is an in vitro method for assessing whether a human subject has periodontitis. The method comprises detecting, in a sample of saliva from said subject, the concentrations of the proteins Hepatocyte Growth Factor (HGF) and Matrix Metalloproteinase 8 (MMP-8). Based on the concentrations determined, and adding age, and possibly other demographic markers such as sex and/or BMI, a testing value reflecting the joint concentrations is determined for said proteins, in combination with one or more demographic markers. The testing value is compared with a threshold value. The threshold reflects in the same manner the joint concentrations and the age, and possibly other demographic markers, as associated with periodontitis and may be seen as an upper limit of testing values as seen in a population of subjects without periodontitis. Thereby a testing value at or above the threshold value is indicative for periodontitis in said subject.

Diagnostics of periodontitis based on salivary HGF and MMP-8

The present invention relates to clinical decision support systems. In detail, the present invention relates to a method for determining the initial fibrinogen concentration in a biological sample, to the use of fibrinogen added to a biological sample in at least one predetermined concentration, to a device for determining the initial fibrinogen concentration in a biological sample, to a computer program comprising program code means for causing a computer to carry out at least several steps of the method according to the invention, to a computer readable non-transitory storage medium containing instructions for carrying out at least some steps of the method according to the invention, and to a kit for determining the initial fibrinogen concentration in a biological sample.

Method for determining the fibrinogen concentration in a biological sample

Presented are concepts for feature identification in medical imaging of a subject. One such concept processes a medical image with a Bayesian deep learning network to determine a first image feature of interest and an associated uncertainty value, the first image feature being located in a first sub-region of the image. It also processes the medical image with a generative adversarial network to determine a second image feature of interest within the first sub-region of the image and an associated uncertainty value. Based on the first and second image features and their associated uncertainty values, the first sub-region of the image is classified.

Feature identification in medical imaging

The present invention relates to device () for determining a fibrinogen level () in a sample () comprising, a first input for obtaining an attenuance signal () over time indicative of a fibrin polymerization of said sample (), a second input for obtaining a reactant concentration signal () over time indicative of a reactant concentration in said sample (), wherein the reactant is any substance leading to the cleavage of fibrinogen to fibrin, a simulation unit () running a model () using the reactant concentration signal () as an input to provide a simulated attenuance signal () over time, and an evaluation unit () configured to infer the fibrinogen level () of said sample () by comparing the attenuance signal () over time with the simulated attenuance signal () over time.

Device, system and method for determining a fibrinogen level in a blood sample

The present invention provides for a simultaneous graphical representation, a risk of bleeding and a risk of thrombosis providing a visualized bridge therapy process. Furthermore, the present invention provides for a computer-based prediction of the haemostatic situation of the examined blood circulation by using a combination of a biochemical model and a pharmacokinetic model for calculation or another mathematical representation of the blood circulation.

Method of visualizing a bridge therapy process

A method for estimating concentrations or concentration changes of particular biological markers, e.g. proteins, includes performing estimations by providing a mathematical model with data from measurements of other biological markers, e.g. other proteins. Thus, instead of measuring the concentration of particular proteins, this concentration can be estimated using the model. The estimated proteins can be used together with other clinical data in another model which is able to provide information or estimations of a patient's disease, for example.

Estimation of diagnostic markers

A method within dynamic molecular imaging comprising dynamically estimating a first parameter (&#x3b2;(x)) and a second parameter (k(x)) of an activity function describing the bio distribution of an administered tracer, is disclosed. More specifically, the method comprises specifying a first target variance (&#x3c3;(x)) and a second target variance (&#x3c3;(x)) of the first parameter (&#x3b2;(x)) and the second parameter (k(x)) of the activity function, respectively; initiating an image acquisition; reconstructing the first parameter (&#x3b2;(x)) and the second parameter (k(x)); reconstructing a first associated variance (&#x3c3;(x)) and a second associated variance (&#x3c3;(x)) of the first parameter (&#x3b2;(x)) and the second parameter (k(x)), respectively; and repeating the image acquisition and the reconstructing until the first associated variance (&#x3c3;(x)) and the second associated variance (&#x3c3;(x)) are equal to or less than the first target variance (&#x3c3;(x)) and the second target variance (&#x3c3;(x)), respectively, &#x3c3;(x)&#x2266;&#x3c3;(x) and &#x3c3;(x)&#x2266;&#x3c3;(x). The method enables patient-specific adaptive protocols within molecular imaging.

Method to individualize kinetic parameter estimation for dynamic molecular imaging procedures

An imaging system () comprises a data device (), which controls radiation data acquisition from a subject positioned in an examination region () for an examination. A rebinning processor () bins the acquired data periodically into a histogram (). A transform () transforms the histogram () into individual independent or uncorrelated components, each component including a signal content and a noise content. A stopping determining device () compares an aspect of at least one selected component to a predetermined threshold (TH) and, based on the comparison, terminates the data acquisition.

Growing community of inventors

Eindhoven, Netherlands

Bart Jacob Bakker