Koninklijke Philips Corporation N.v.

Asher Gringauz

Liran Goshen

Yechiel Lamash

Kevin Martin Brown

Jens Wiegert

Stanislav Zabic

Jonathan Lessick

Andrei Feldman

Yoav Bar

Guido Pardo-Roques

Jonathan Sapir

A mechanism for processing input 3D image data. In a first phase, the input 3D image data is separately processed using one or more neural networks to produce one or more modified 3D image data. In a second phase, the input 3D image data and the modified 3D image data are processed using neural networks to produce an output. The neural networks that produce the modified 3D image data are configured to process slices or sub-volumes of the input 3D image data to produce modified 3D image data.

Network architecture for 3D image processing

A method includes obtaining original image data and obtaining de-noised image data, wherein the de-noised image data is the original image data after de-noising the original image data. The method further includes determining a noise pattern for the original image data. The method further includes estimating underlying local structure from the original image data based on the noise pattern. The method further includes restoring low contrast structure that is lost during the de-noising of the original image data to the de-noised image data based on the estimated underlying local structure by adding the estimated underlying local structure to the de-noise imaged data, generating low contrast structure enhanced de-noised image data.

Restoration of low contrast structure in de-noise image data

A method for segmenting bone in spectral image data is described herein. The spectral image data includes at least a first set of image data corresponding to a first energy and second set of image data corresponding to a second different energy. The method includes obtaining the spectral image data. The method further includes extracting a set of features for each voxel in spectral image data. The method further includes determining, for each voxel, a probability that each voxel represents bone structure based on the set of features. The method further includes extracting bone structure from the spectral image data based on the probabilities.

Bone segmentation from image data

A method for improving image quality of image data includes analyzing, for each of a plurality of voxels of image data, a set of entries of a dictionary, wherein an entry represents a mapping between a lower resolution patch of voxels and a corresponding higher resolution patch of voxel or a local neighborhood around a voxel, deriving, for each of the plurality of voxels, a subspace based on the analysis, wherein the subspace is for one of the mapping or the local neighborhood, and restoring target image data based on the subspaces, wherein the target image data is image data with higher image resolution or reduced image noise.

Image noise reduction and/or image resolution improvement

A method includes enhancing a contrast to noise ratio (CNR) of image data, generating CNR enhanced image data, wherein the CNR enhanced image data has a substantially same image quality as the image data. A computing system () includes a computer readable storage medium () encoded with computer readable instructions for enhancing a contrast to noise ratio (CNR) of image data and one or more processors (), which, when executing the computer readable instructions, causes the computing system to enhance the CNR of the image data. A method includes generating CNR enhanced image data, wherein CNR enhanced image data has a substantially same noise level, noise power spectrum, and spatial resolution of the image data.

Contrast to noise ratio (CNR) enhancer

A method includes generating enhanced image data based on lower dose image data and a predetermined image quality threshold, wherein an image quality of the enhanced image data is substantially similar to an image quality of higher dose image data, and a system includes an image quality enhancer () that generates enhanced image data based on lower dose image data and a predetermined image quality threshold, wherein an image quality of the enhanced image data is substantially similar to an image quality of higher dose image data.

Enhanced image data/dose reduction

A method for classifying tissue as normal or abnormal tissue includes obtaining segmented reconstructed volumetric image data for predetermined tissue of interest, generating a 2D voxel representation of the segmented reconstructed volumetric image data, and classifying voxels of the segmented reconstructed volumetric image data as corresponding to abnormal and normal tissue based on the 2D voxel representation.

Tissue classification

A method for artifact reduction in CT images comprising reconstructing a first data image using an inexact reconstruction algorithm, segmenting the first data image to provide a second image with high attenuation objects separate from low attenuation objects, reprojecting the second image to form a second set of data, reconstructing a third image from the second data set using an inexact reconstruction algorithm; and subtracting at least those portions of the third image outside the high attenuation object from the first image.

Growing community of inventors

Nesher, Israel

Asher Gringauz