Medical Image Segmentation

Simple SummaryContent extracted from patent full text and abstract with AI.

This invention describes an advanced system and method for segmenting medical images using neural networks. The system uses two (or more) neural networks (controller networks) that process different sequences of image patches and share a memory that allows them to exchange context information. This shared memory architecture enables the networks to better utilize global anatomical context, resulting in more accurate and efficient segmentation of organs or other structures in 2D, 3D, or even 4D medical images such as MRI or CT scans.

Use CasesContent extracted from patent full text and abstract with AI.

• Automated medical image segmentation (e.g., MRI, CT, X-ray, ultrasound) to identify organs, tumors, or other structures.
• Computer-aided diagnosis where accurate segmentation is needed for disease detection and treatment planning.
• Pre-processing step for medical image analysis, quantitative measurement, or 3D reconstruction.
• Assisting radiologists in annotation and analysis by speeding up image review.
• Enabling real-time segmentation during surgical or interventional imaging workflows.
• Segmentation for research purposes in clinical trials and computational anatomy studies.

BenefitsContent extracted from patent full text and abstract with AI.

• Improves segmentation accuracy by leveraging global anatomical context through shared neural memory.
• Provides highly efficient and reliable segmentation even on high-dimensional (e.g., 3D/4D) images.
• Enables robust and fast automated segmentation, reducing manual work for clinicians.
• Scalable and adaptable to different types of medical imaging modalities (MRI, CT, ultrasound, etc.).
• Architecture supports easy training and updating with new data and anatomical knowledge.
• Facilitates real-time or near real-time segmentation in clinical and research settings.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

The invention provides systems and methods for medical image processing using neural networks. A first and a second controller network share a memory to which both the first and second controller network can write data and from which both the first and the second controller network can read data. Reading and writing is performed by respective read and write heads which are advantageously neural networks trained how to write and read in an optimal way. The memory thus provides each controller network with context data generated by the respective other controller network.