Publications by authors named "Sambit Satpathy"
Shape Memory Alloys (SMAs) are pivotal in diverse industrial applications due to their exceptional properties, including actuation, biocompatibility, and adaptability in aerospace, biomedical, and military domains. However, their complex machinability often leads to high costs and suboptimal surface quality when processed using traditional methods. Using Response Surface Methodology (RSM) with a Central Composite Design (CCD), this study evaluated the effects of input parameters, including pulse on time (T), pulse off time (T), peak current (Ip), and gap voltage (GV), on material wear responses during Electrical Discharge Machining (EDM).
View Article and Find Full Text PDFThis paper introduces 3D-QTRNet, a novel quantum-inspired neural network for volumetric medical image segmentation. Unlike conventional CNNs, which suffer from slow convergence and high complexity, and QINNs, which are limited to grayscale segmentation, our approach leverages qutrit encoding and tensor ring decomposition. These techniques improve segmentation accuracy, optimize memory usage, and accelerate model convergence.
View Article and Find Full Text PDFArticle Synopsis
- - The work focuses on
- three key innovations
- : smart data collection, an optimized training algorithm, and a
- Bayesian approach combined with split learning
- to ensure the privacy of patent data.
- - By utilizing consumer electronics like
- wearable devices
- and the Internet of Things (IoT) to capture THz images, the researchers implement an
- EM algorithm for training
- using a new split learning method aimed at improving imaging depth and tissue contrast for early breast cancer detection.
- - Their
- hybrid algorithm achieves an impressive accuracy of 97.5%
- over 100 training epochs, outperforming older models that required more training epochs (e.g., 165) for less accuracy.