Health

Achievement in Biomedical Engineering Automatic 3D Brain Tumor Segmentation using Deep Learning

December 28, 2022

Sandeep Trivedi is a Senior SAP Analytics Consultant at Deloitte LLP in Houston, Texas, United States. He received his Bachelor’s degree in Electronic & Communication Engineering from Rajiv Gandhi Proudyogiki Vishwavidyalaya, Bhopal, India, in 2005. He is SAP certified and specializes in SAP Analytics Reporting using machine learning algorithms. He is not only an SAP Analytics Consultant but also a visionary, researcher, and technical committee member of various prestigious international conferences.

He participated in and led multiple SAP technology-related projects in India, Qatar, UAE, Malaysia, Singapore, and the United States. He contributes to engaging students in STEM conversation at NASA and supports TESI staff in this regard. He is a Senior Member of IEEE, U.S., Region 5, and the American SAP User Group (ASUG). Prestigious IEEE conferences, including ICCCIS-2022, IEEE UEMCON 2022, and IEEE ICSCA 2022, invite Sandeep Trivedi as a technical reviewer because of his expertise in Artificial Intelligence, Machine Learning, and Deep Learning. IEEE Recently featured Sandeep’s article on its official website.

3D structures instead of 2D images reveal more information about brain tumors. Physicians find it more beneficial to diagnose the brain tumor from the 3D structures. However, segmenting the tumors from Computer Tomography (CT) images and reconstructing the 3D shape is complicated and sensitive. As a matter of fact, anything related to healthcare is sensitive.

A slight mistake in 3D brain tumor segmentation and reconstruction may lead to improper diagnosis causing severe health and emotional damage to the patients. Sandeep Trivedi has developed an innovative 3D brain tumor segmentation method that will be a breakthrough in biomedical engineering.

The process uses a Nested Deep Neural Network architecture. Its purpose is to automate the brain tumor diagnosis process. This innovative technology will be integrated with modern CT scanners to assist radiologists. Brain tumors are sensitive. The more information the physicians get about the tumor, the more effective treatment they can provide.

A surgical procedure is mandatory to remove the tumor if not curable through medicine. Imagine how helpful it would be for brain surgeons to have the brain tumor 3D printed before they start the surgical procedure. It would give the surgeons a crystal-clear idea of what they are going to surgery to improve the success rate.

Sandeep Trivedi

Brain surgery is critical, and complications may arise anytime during the surgery. The 3D-printed tumor would help the surgeons assess possible challenges and take proper precautions before they start the surgery. From this vision, Sandeep Trivedi developed the segmentation method, which has been published in the 13th IEEE Annual Ubiquitous Computing, Electronics, and Mobile Communication Conference (UEMCON 2022). The system segments brain tumors in 3D with 90.2%, 85.11%, and 85.41% dice scores for whole tumor, core tumor, and enhancing tumor, respectively.

Sandeep envisions integrating his innovation with the CT scanner along with a 3D printing module to 3D print the tumor after scanning. It sounds futuristic. There are many challenges associated with it. It requires more research and experiments. Sandeep Trivedi believes he can turn this futuristic idea into a realistic solution.