Back to results
US 12,731,420 B2Granted
Method and system for prediction of microorganism growth using artificial intelligence
AssigneeBacteromic Sp. Z.O.O.
InventorsPawel Debski, Krystian Kurek, Michal Bortkiewicz, Justyna Szeluga
Life of this patent
Estimating the term…
Abstract
Disclosed is a computer-implemented method for real-time prediction of microorganism growth or microorganism growth inhibition of a microorganism inoculum in a phenotypic antimicrobial susceptibility test (AST) as well as a rapid antimicrobial susceptibility testing system respectively using artificial intelligence.
Claims19
A computer-implemented method for real-time prediction of microorganism growth or microorganism growth inhibition of a microorganism inoculum in a phenotypic antimicrobial susceptibility test (AST), the method comprising the following steps:
- a) Incubating a sample of a microorganism with a suitable incubating device, wherein the incubated sample comprises a single antimicrobial agent or a combination of antimicrobial agents,
- b) Taking a sequence of two or more digital images of the incubated sample microorganism of step a) with a suitable imaging device and providing information on pixel intensity (pi) per pixel of the two or more digital images, wherein the sequence of the two or more digital images is consecutively distributed within a progress value (tp) with 0<tp<1, wherein the progress value tp represents a ratio of a proportional incubation time period for the consecutively distributed sequence of the two or more digital images of the incubated sample microorganism divided by an overall incubation time period from start to end of the incubation for the respective microorganism, and
- c) Predicting microorganism growth or microorganism growth inhibition for the incubated sample microorganism as a function of an output score (So) 0≤So≤1 of a deep learning neural network configured to extract spatio-temporal features of the sequence of the two or more digital images and to classify the respective images, wherein So=0 represents a distinct microorganism growth inhibition and So=1 represents a distinct microorganism growth, and wherein the deep learning neural network uses as input data:
- (i) a sequence of pixel intensity (pi) per pixel for the two or more digital images of the incubated sample microorganism of step b), or
- (ii) a determined sequence of differences in the pixel intensity (Δpi) per pixel between two consecutive images respectively for the two or more digital images of step b), or
- (iii) a sequence of pixel intensity (pi) per pixel for the two or more digital images of the incubated sample microorganism of step b) and a determined sequence of differences in the pixel intensity (Δpi) per pixel between two consecutive images respectively for the two or more digital images of step b).
3 more elements
A rapid antimicrobial susceptibility testing system for predicting microorganism growth or microorganism growth inhibition of a microorganism inoculum in a phenotypic antimicrobial susceptibility test (AST), the system comprising:
- a) an incubation assembly configured for housing an incubation device for incubating a sample of a microorganism, wherein the sample comprises a single antimicrobial agent or a combination of antimicrobial agents,
- b) an imaging device configured to taking a sequence of two or more digital images of the incubated sample microorganism of feature a) on the incubating device, wherein the imaging device is configured to provide information on pixel intensity (pi) per pixel for each of the two or more digital images and wherein the imaging device is further configured to taking the sequence of the two or more digital images consecutively distributed within a progress value (tp) with 0<tp<1, wherein the progress value tp represents a ratio of a proportional incubation time period for the consecutively distributed sequence of the two or more digital images of the incubated sample microorganism divided by an overall incubation time period from start to end of the incubation for the respective microorganism, and
- c) a computer assembly comprising one or more processors, and an analysis module comprising a deep neural network unit configured to extract spatio-temporal features of the sequence of the two or more digital images and to classify the two or more digital images and one or more computer-readable media storing instructions that, when executed by the one or more processors, cause the one or more processors to perform predicting microorganism growth or microorganism growth inhibition for the incubated sample microorganism as a function of an output score (So) 0≤So≤1 of the deep learning neural network, wherein So=0 represents a distinct microorganism growth inhibition and So=1 represents a distinct microorganism growth, and wherein the deep learning neural network is configured to use as input data:
- (i) a sequence of pixel intensity (pi) per pixel for the two or more images of the incubated sample microorganism of feature b), or
- (ii) a determined sequence of differences in the pixel intensity (Δpi) per pixel between two consecutive images respectively for the two or more images of feature b), or
- (iii) a sequence of pixel intensity (pi) per pixel for the two or more images of the incubated sample microorganism of feature b) and a determined sequence of differences in the pixel intensity (Δpi) per pixel between two consecutive images respectively for the two or more images of feature b).
3 more elements
Dependencies are inferred from the claim text.