Ge Precision Healthcare LLC

General Electric Company

Sudhanya Chatterjee

Dattesh Dayanand Shanbhag

Rohan Keshav Patil

Rakesh Mullick

Krishna Seetharam Shriram

Mahendra Madhukar Patil

Florian Wiesinger

Kevin Matthew Koch

Deepa Anand

Sheshadri Thiruvenkadam

Florintina C

Syed Asad Hashmi

Suresh Joel

Suresh Emmanuel Devadoss Joel

Megha Goel

Preetham Shankapal

The disclosure relates to multiplanar reformation of three-dimensional medical images. In particular, the invention provides a method for reformatting image sequences by determining a landmark plane intersecting a volume, acquiring an image sequence, reformatting the image sequence along the landmark plane to produce a first reformatted image sequence, perturbing the landmark plane to produce a perturbed landmark plane, reformatting the first reformatted image sequence along the perturbed landmark plane to produce a second reformatted image sequence, mapping the second reformatted image sequence, the image sequence, and the landmark plane, to a resolution enhanced image sequence using a trained image enhancement network, and displaying the resolution enhanced image sequence via a display device. The present disclosure provides approaches which may reduce image artifacts in retrospectively reformatted image sequences, particularly in cases of retrospective reformatting of medium or low-resolution image sequences, without relying on acquisition of high-resolution 3D images.

Methods and systems for multiplanar reformation with machine learning based image enhancement

Systems/techniques that facilitate multi-modal image registration via modality-neutral machine learning transformation are provided. In various embodiments, a system can access a first image and a second image, where the first image can depict an anatomical structure according to a first imaging modality, and where the second image can depict the anatomical structure according to a second imaging modality that is different from the first imaging modality. In various aspects, the system can generate, via execution of a machine learning model on the first image and the second image, a modality-neutral version of the first image and a modality-neutral version of the second image. In various instances, the system can register the first image with the second image, based on the modality-neutral version of the first image and the modality-neutral version of the second image.

Multi-modal image registration via modality-neutral machine learning transformation

Techniques are described for generating high quality training data collections for training artificial intelligence (AI) models in the medical imaging domain. A method embodiment comprises receiving, by a system comprising processor, input indicating a clinical context associated with usage of a medical image dataset, and selecting, by the system, one or more data scrutiny metrics for filtering the medical image dataset based on the clinical context. The method further comprises applying, by the system, one or more image processing functions to the medical image dataset to generate metric values of the one or more data scrutiny metrics for respective medical images included in the medical image dataset, filtering, by the system, the medical image dataset into one or more subsets based on one or more acceptability criteria for the metric values.

Generating high quality training data collections for training artificial intelligence models

Various methods and systems are provided for generating super-resolution images. In one embodiment, a method comprises: progressively up-sampling an input image to generate a super-resolution output image by: generating N intermediate images based on the input image, where N is equal to at least one, including a first intermediate image by providing the input image to a deep neural network, where a resolution of the first intermediate image is a multiple of a resolution of the input image, higher than the resolution of the input image, and can be any positive real value and not necessarily an integer value; generating the super-resolution output image based on the N intermediate images, the super-resolution output image having a resolution higher than a respective resolution of each intermediate image of the N intermediate images and the resolution of the input image; and displaying the super-resolution output image via a display device.

Methods and systems for super-resolution with progressive sub-voxel up-sampling

Systems and methods are provided for reconstructing images from motion-affected k-space data. In one example, a method comprises obtaining k-space data of a spin echo magnetic resonance imaging (MRI) exam of a subject, the k-space data comprising a plurality of echo train lengths (ETLs), with each ETL comprising a subset of lines of the k-space data. The method further comprises identifying a subset of ETLs of the plurality of ETLs of the k-space data corresponding to a dominant pose of the subject, generating an undersampled version of the k-space data, the undersampled version including only the subset of ETLs, entering the undersampled version of the k-space data as input to a reconstruction model trained to output a reconstructed image based on the undersampled version of the k-space data, and displaying the reconstructed image on a display device and/or saving the reconstructed image in memory.

Systems and methods for motion-artifact reduction in magnetic resonance imaging

Systems and methods for deep learning based magnetic resonance imaging (MRI) examination acceleration are provided. The method of deep learning (DL) based magnetic resonance imaging (MRI) examination acceleration comprises acquiring at least one fully sampled reference k-space data of a subject and acquiring a plurality of partial k-space of the subject. The method further comprises grafting the plurality of partial k-space with the at least one fully sampled reference k-space data to generate a grafted k-space for accelerated examination. The method further comprises training a deep learning (DL) module using the fully sampled reference k-space data and the grafted k-space to remove the grafting artifacts.

Deep learning based magnetic resonance imaging (MRI) examination acceleration

A computer-implemented method for generating an artifact corrected reconstructed contrast image from magnetic resonance imaging (MRI) data is provided. The method includes inputting into a trained deep neural network both a synthesized contrast image derived from multi-delay multi-echo (MDME) scan data or the MDME scan data acquired during a first scan of an object of interest utilizing a MDME sequence and a composite image, wherein the composite image is derived from both the MDME scan data and contrast scan data acquired during a second scan of the object of interest utilizing a contrast MRI sequence. The method also includes utilizing the trained deep neural network to generate the artifact corrected reconstructed contrast image based on both the synthesized contrast image or the MDME scan data and the composite image. The method further includes outputting from the trained deep neural network the artifact corrected reconstructed contrast image.

System and method for deep learning-based generation of true contrast images utilizing synthetic magnetic resonance imaging data

A medical imaging system having at least one medical imaging device providing image data of a subject is provided. The medical imaging system further includes a processing system programmed to train a deep learning (DL) network using a plurality of training images to predict noise in input data. The plurality of training images includes a plurality of excitation (NEX) images acquired for each line of k-space training data. The processing system is further programmed to use the trained DL network to determine noise in the image data of the subject and to generate a denoised medical image of the subject having reduced noise based on the determined noise in the image data.

Deep learning based medical system and method for image acquisition

An imaging system and method are disclosed. An MR image and measured B0 field map of a target volume in a subject are reconstructed, where the MR image includes one or more bright and/or dark regions. One or more distinctive constituent materials corresponding to the bright regions are identified. Each dark region is iteratively labeled as one or more ambiguous constituent materials. Susceptibility values corresponding to each distinctive and iteratively labeled ambiguous constituent material is assigned. A simulated B0 field map is iteratively generated based on the assigned susceptibility values. A similarity metric is determined between the measured and simulated B0 field maps. Constituent materials are identified in the dark regions based on the similarity metric to ascertain corresponding susceptibility values. The MRI data is corrected based on the assigned and ascertained susceptibility values. A diagnostic assessment of the target volume is determined based on the corrected MRI data.

Growing community of inventors

Bengaluru, India

Sudhanya Chatterjee