View source for 20240005512. METHODS AND SYSTEMS FOR AUTOMATED IMAGE SEGMENTATION OF ANATOMICAL STRUCTURE simplified abstract (Tata Consultancy Services Limited)

=METHODS AND SYSTEMS FOR AUTOMATED IMAGE SEGMENTATION OF ANATOMICAL STRUCTURE=


==Organization Name==

[[:Category:Tata Consultancy Services Limited|Tata Consultancy Services Limited]]


[[Category:Tata Consultancy Services Limited]]

==Inventor(s)==

[[:Category:Aparna Kanakatte Gurumurthy of Bangalore (IN)|Aparna Kanakatte Gurumurthy of Bangalore (IN)]][[Category:Aparna Kanakatte Gurumurthy of Bangalore (IN)]]

[[:Category:Avik Ghose of Kolkata (IN)|Avik Ghose of Kolkata (IN)]][[Category:Avik Ghose of Kolkata (IN)]]

[[:Category:Divya Manoharlal Bhatia of Bangalore (IN)|Divya Manoharlal Bhatia of Bangalore (IN)]][[Category:Divya Manoharlal Bhatia of Bangalore (IN)]]

[[:Category:Jayavardhana Rama Gubbi Lakshminarasimha of Bangalore (IN)|Jayavardhana Rama Gubbi Lakshminarasimha of Bangalore (IN)]][[Category:Jayavardhana Rama Gubbi Lakshminarasimha of Bangalore (IN)]]

==METHODS AND SYSTEMS FOR AUTOMATED IMAGE SEGMENTATION OF ANATOMICAL STRUCTURE - A simplified explanation of the abstract==


This abstract first appeared for US patent application 20240005512 titled 'METHODS AND SYSTEMS FOR AUTOMATED IMAGE SEGMENTATION OF ANATOMICAL STRUCTURE

==Simplified Explanation==

The patent application is about a new method and system for automated 3D image segmentation of anatomical structures, specifically the heart. The existing techniques in the literature mainly use 2D or slice-by-slice data for training, which lacks 3D contextual information and is inaccurate and inefficient in the segmentation of the last slice of the image. 

The innovation proposed in this patent application is a new architecture based on generative adversarial networks (GAN) for 3D segmentation. It utilizes a patch-based extraction technique and a class-weighted generalized dice loss to improve accuracy. This architecture is capable of storing and utilizing 3D contextual information for more accurate segmentation of anatomical structures like the heart.

* New method and system for automated 3D image segmentation of anatomical structures
* Uses a generative adversarial network (GAN) based architecture
* Incorporates a patch-based extraction technique and a class-weighted generalized dice loss
* Improves accuracy and efficiency in 3D image segmentation
* Stores and utilizes 3D contextual information for more accurate segmentation

==Potential Applications==
This technology has potential applications in various medical fields where accurate 3D image segmentation of anatomical structures is required. Some potential applications include:

* Cardiology: Accurate segmentation of the heart for diagnosis and treatment planning
* Radiology: Precise segmentation of organs and tissues for medical imaging analysis
* Oncology: Improved segmentation of tumors for radiation therapy planning
* Neurology: Accurate segmentation of brain structures for diagnosis and research purposes

==Problems Solved==
The technology solves several problems associated with automated 3D image segmentation of anatomical structures:

* Lack of 3D contextual information in existing techniques
* Inaccuracy and inefficiency in segmentation of the last slice of the image
* Limited accuracy and efficiency in conventional 3D image segmentation methods

==Benefits==
The benefits of this technology are:

* Improved accuracy in 3D image segmentation of anatomical structures
* Efficient and reliable segmentation of the last slice of the image
* Utilization of 3D contextual information for more accurate segmentation
* Potential for better diagnosis, treatment planning, and medical research in various fields


==Original Abstract Submitted==

this disclosure relates generally to methods and systems for automated image segmentation of an anatomical structure such as heart. most of the techniques in literature are using 2-d or slice by-slice data due to lightweight and need of less data for training. these networks lack 3-d contextual information. further, the conventional techniques are inaccurate and inefficient in the 3-d image segmentation till the last slice of the image. the present disclosure solves automated 3-d image segmentation of the anatomical structure such as heart, by proposing a new generative adversarial network (gan) based architecture for the 3-d segmentation, with a patch-based extraction technique and a class-weighted generalized dice loss. the proposed 3-d gan based architecture is capable of storing the 3-d contextual information for the image segmentation of the anatomical structure, with high accuracy.



[[Category:G06T7/12]]
[[Category:G06T15/00]]
[[Category:G06V10/25]]
[[Category:G06V10/44]]
[[Category:G06V10/764]]
[[Category:G06V10/771]]