IEEE SPS SBC Webinar: Data-driven Approaches for Improved Reconstruction and Segmentation in Medical Imaging.(By Dr. Raji Susan Mathew)
This talk deals with two aspects of medical Imaging: image reconstruction and segmentation. The first part covers the image reconstruction for magnetic resonance imaging (MRI). Despite the capability of providing high-resolution images, the difficulties associated with lengthy acquisition time necessitate reconstruction of the final image from a limited number of k-space samples. The reconstruction problem falls under the broad class of ill-posed inverse problems, and regularization is necessary for obtaining stable and meaningful solutions. However, the accuracy of regularized output depends on the regularization parameter choice. The adaptive estimation of the regularization parameter from the data for sparsity-promoting methods will be discussed. The second part focuses on nerve segmentation in ultrasound images. The automated segmentation of the median nerve at the wrist and from wrist to elbow using different deep learning models along with the associated challenges will be discussed. Finally, the talk will conclude with a discussion on the implementation of the model in real-time.
Date and Time
Location
Hosts
Registration
- Date: 11 Oct 2023
- Time: 07:00 PM to 08:00 PM
- All times are (UTC+05:30) Chennai
-
Add Event to Calendar
Speakers
Dr. Raji Susan Mathew
Data-driven Approaches for Improved Reconstruction and Segmentation in Medical Imaging.
This talk deals with two aspects of medical Imaging: image reconstruction and segmentation. The first part covers the image reconstruction for magnetic resonance imaging (MRI). Despite the capability of providing high-resolution images, the difficulties associated with lengthy acquisition time necessitate reconstruction of the final image from a limited number of k-space samples. The reconstruction problem falls under the broad class of ill-posed inverse problems, and regularization is necessary for obtaining stable and meaningful solutions. However, the accuracy of regularized output depends on the regularization parameter choice. The adaptive estimation of the regularization parameter from the data for sparsity-promoting methods will be discussed. The second part focuses on nerve segmentation in ultrasound images. The automated segmentation of the median nerve at the wrist and from wrist to elbow using different deep learning models along with the associated challenges will be discussed. Finally, the talk will conclude with a discussion on the implementation of the model in real-time.
Biography:
Dr. Raji Susan Mathew is a C. V. Raman Post-doctoral Fellow in the Medical Imaging Group, headed by Prof. Phaneendra K. Yalavarthy in the Department of Computational and Data Sciences IISc Bangalore, since 2021. She completed Ph.D. in the area of MR image reconstruction from Indian Institute of Information Technology and Management- Kerala under the guidance of Prof. Joseph Suresh Paul. Her Ph.D. work was focused on the proper use of regularised solution that facilitates noise reduction and artefact suppression in the reconstructed magnetic resonance image. Prior to Ph.D she received bachelor’s degree in Electronics and Communication Engineering from the Mahatma Gandhi university, Kottayam and master's degree in Signal Processing from the Cochin university of science and technology, Kochi in 2011 and 2013 respectively. Her research interests include application of regularisation for image reconstruction techniques, compressed sensing and deep learning. She is also interested in inverse problems and computational methods for medical imaging.
Email:
Address: IISc Bangalore, , Bangalore, India