Next-Generation Synthetic Breast Tissue Image Generation System
Legal Citation
Summary of the Inventive Concept
A novel system and method for generating synthetic breast tissue images with enhanced tissue visibility, utilizing advanced machine learning and physics-based models to overcome limitations of traditional high-density element suppression techniques.
Background and Problem Solved
Traditional breast imaging techniques, such as tomosynthesis, often struggle with obtrusive high-density elements, which can obscure critical tissue structures. The original patent addressed this issue through high-density element suppression, but limitations remain. The new inventive concept addresses these limitations by introducing advanced machine learning and physics-based models to generate synthetic breast tissue images with enhanced tissue visibility.
Detailed Description of the Inventive Concept
The next-generation system comprises a neural network trained on a dataset of high-density element-suppressed images to predict suppressed tissue structures. The system also includes a processing module to generate a synthesized image with enhanced tissue visibility. Alternatively, a generative adversarial network can be trained on paired high-density element-suppressed and corresponding high-density element-enhanced images to generate a synthesized image. The system can also utilize a database of high-density element-suppressed images and a machine learning model to predict suppressed tissue structures. Additionally, a high-density element detection module can identify regions of high-density elements in an input image, and a processing module can apply a physics-based model to simulate the suppression of high-density elements in the identified regions.
Novelty and Inventive Step
The new inventive concept introduces the use of advanced machine learning models, such as neural networks and generative adversarial networks, to predict suppressed tissue structures and generate synthetic breast tissue images with enhanced tissue visibility. This represents a significant departure from traditional high-density element suppression techniques, offering improved image quality and diagnostic accuracy.
Alternative Embodiments and Variations
Alternative embodiments may include the use of other machine learning models, such as deep learning-based image enhancement algorithms, or the integration of additional data sources, such as patient-specific information or multi-modal imaging data. Variations may also include the application of the inventive concept to other medical imaging modalities, such as MRI or CT scans.
Potential Commercial Applications and Market
The next-generation synthetic breast tissue image generation system has significant commercial potential in the medical imaging industry, particularly in the areas of breast cancer screening and diagnosis. The system's ability to generate high-quality images with enhanced tissue visibility can improve diagnostic accuracy and patient outcomes, making it an attractive solution for healthcare providers and medical device manufacturers.
CPC Classifications
| Section | Class | Group |
|---|---|---|
| A | A61 | A61B6/5258 |
| A | A61 | A61B6/461 |
| A | A61 | A61B6/502 |
| G | G06 | G06T11/008 |
| G | G06 | G06T2207/10116 |
| G | G06 | G06T2207/30068 |
Section 103 Obviousness Analysis (PHOSITA)
Field of Art
Medical imaging, specifically breast tissue tomosynthesis and synthetic image generation, involving advanced image processing techniques, machine learning, and radiological image enhancement technologies
Person of Ordinary Skill (PHOSITA) Profile
A skilled practitioner with expertise in medical imaging, computer vision, machine learning, and radiological image processing, typically holding a PhD or equivalent professional experience in biomedical engineering, medical imaging, or computational radiology
Obviousness Rationale
The published technical disclosure represents predictable variations on the source patent's core methodology of synthetic breast tissue image generation by applying well-established machine learning techniques to the existing high-density element suppression framework. A person having ordinary skill would recognize that neural networks and generative adversarial networks are standard approaches for image enhancement and reconstruction tasks in medical imaging. The proposed variations merely substitute traditional image processing algorithms with contemporary machine learning models to achieve substantially similar outcomes of improved tissue visibility.
Obvious Combinations & Variations
Original Patent Information
| Patent Number | US 11,857,358 |
|---|---|
| Title | System and method for synthetic breast tissue image generation by high density element suppression |
| Assignee(s) | Hologic, Inc. |