Publications by authors named "Elizabeth C Behrman"
Qutrit-Inspired Fully Self-Supervised Shallow Quantum Learning Network for Brain Tumor Segmentation.
IEEE Trans Neural Netw Learn Syst
November 2022
Article Synopsis
- Classical self-supervised networks struggle with convergence and segmentation accuracy, prompting the development of a new model called the quantum fully self-supervised neural network (QFS-Net) for brain MR image segmentation.
- The QFS-Net utilizes a three-level qutrit quantum information system and a unique layered structure interconnected by Hadamard gates, allowing it to perform faster, unsupervised learning.
- Tested on the Cancer Imaging Archive dataset, QFS-Net showed superior performance in detecting tumors compared to existing models like U-Net and URes-Net, requiring less human involvement and computational power, while also displaying strong results on natural gray-scale images.