-
FedCL-Ensemble Learning: A Framework of Federated Continual Learning with Ensemble Transfer Learning Enhanced for Alzheimer's MRI Classifications while Preserving Privacy
Authors:
Rishit Kapoor,
Jesher Joshua,
Muralidharan Vijayarangan,
Natarajan B
Abstract:
This research work introduces a novel approach to the classification of Alzheimer's disease by using the advanced deep learning techniques combined with secure data processing methods. This research work primary uses transfer learning models such as ResNet, ImageNet, and VNet to extract high-level features from medical image data. Thereafter, these pre-trained models were fine-tuned for Alzheimer'…
▽ More
This research work introduces a novel approach to the classification of Alzheimer's disease by using the advanced deep learning techniques combined with secure data processing methods. This research work primary uses transfer learning models such as ResNet, ImageNet, and VNet to extract high-level features from medical image data. Thereafter, these pre-trained models were fine-tuned for Alzheimer's related subtle patterns such that the model is capable of robust feature extraction over varying data sources. Further, the federated learning approaches were incorporated to tackle a few other challenges related to classification, aimed to provide better prediction performance and protect data privacy. The proposed model was built using federated learning without sharing sensitive patient data. This way, the decentralized model benefits from the large and diversified dataset that it is trained upon while ensuring confidentiality. The cipher-based encryption mechanism is added that allows us to secure the transportation of data and further ensure the privacy and integrity of patient information throughout training and classification. The results of the experiments not only help to improve the accuracy of the classification of Alzheimer's but at the same time provides a framework for secure and collaborative analysis of health care data.
△ Less
Submitted 15 November, 2024;
originally announced November 2024.
-
Feedback Linearizable Discretizations of Second Order Mechanical Systems using Retraction Maps
Authors:
Shreyas N. B.,
David Martin Diego,
Ravi Banavar
Abstract:
Mechanical systems are most often described by a set of continuous-time, nonlinear, second-order differential equations (SODEs) of a particular structure governed by the covariant derivative. The digital implementation of controllers for such systems requires a discrete model of the system and hence requires numerical discretization schemes. Feedback linearizability of such sampled systems, howeve…
▽ More
Mechanical systems are most often described by a set of continuous-time, nonlinear, second-order differential equations (SODEs) of a particular structure governed by the covariant derivative. The digital implementation of controllers for such systems requires a discrete model of the system and hence requires numerical discretization schemes. Feedback linearizability of such sampled systems, however, depends on the discretization scheme employed. In this article, we utilize retraction maps and their lifts to construct feedback linearizable discretizations for SODEs which can be applied to many mechanical systems.
△ Less
Submitted 3 October, 2024; v1 submitted 27 March, 2024;
originally announced March 2024.
-
Robust and Scalable Techniques for TWR and TDoA based localization using Ultra Wide Band Radios
Authors:
Rakshit Ramesh,
Aaron John-Sabu,
Harshitha S,
Siddarth Ramesh,
Vishwas Navada B,
Mukunth Arunachalam,
Bharadwaj Amrutur
Abstract:
Current trends in autonomous vehicles and their applications indicates an increasing need in positioning at low battery and compute cost. Lidars provide accurate localization at the cost of high compute and power consumption which could be detrimental for drones. Modern requirements for autonomous drones such as No-Permit-No-Takeoff (NPNT) and applications restricting drones to a corridor require…
▽ More
Current trends in autonomous vehicles and their applications indicates an increasing need in positioning at low battery and compute cost. Lidars provide accurate localization at the cost of high compute and power consumption which could be detrimental for drones. Modern requirements for autonomous drones such as No-Permit-No-Takeoff (NPNT) and applications restricting drones to a corridor require the infrastructure to constantly determine the location of the drone. Ultra Wide Band Radios (UWB) fulfill such requirements and offer high precision localization and fast position update rates at a fraction of the cost and battery consumption as compared to lidars and also have greater network availability than GPS in a dense forested campus or an indoor setting. We present in this paper a novel protocol and technique to localize a drone for such applications using a Time Difference of Arrival (TDoA) approach. This further increases the position update rates without sacrificing on accuracy and compare it to traditional methods
△ Less
Submitted 10 August, 2020;
originally announced August 2020.
-
CLCNet: Deep learning-based Noise Reduction for Hearing Aids using Complex Linear Coding
Authors:
Hendrik Schröter,
Tobias Rosenkranz,
Alberto N. Escalante B.,
Marc Aubreville,
Andreas Maier
Abstract:
Noise reduction is an important part of modern hearing aids and is included in most commercially available devices. Deep learning-based state-of-the-art algorithms, however, either do not consider real-time and frequency resolution constrains or result in poor quality under very noisy conditions. To improve monaural speech enhancement in noisy environments, we propose CLCNet, a framework based on…
▽ More
Noise reduction is an important part of modern hearing aids and is included in most commercially available devices. Deep learning-based state-of-the-art algorithms, however, either do not consider real-time and frequency resolution constrains or result in poor quality under very noisy conditions. To improve monaural speech enhancement in noisy environments, we propose CLCNet, a framework based on complex valued linear coding. First, we define complex linear coding (CLC) motivated by linear predictive coding (LPC) that is applied in the complex frequency domain. Second, we propose a framework that incorporates complex spectrogram input and coefficient output. Third, we define a parametric normalization for complex valued spectrograms that complies with low-latency and on-line processing. Our CLCNet was evaluated on a mixture of the EUROM database and a real-world noise dataset recorded with hearing aids and compared to traditional real-valued Wiener-Filter gains.
△ Less
Submitted 28 January, 2020;
originally announced January 2020.
