No abstract

Diabetic retinopathy (DR) is a disease resulting from diabetes complications, causing non-reversible damage to retina blood vessels. DR is a leading cause of blindness if not detected early. The currently available DR treatments are limited to stopping or delaying the deterioration of sight, highlighting the importance of regular scanning using high-efficiency computer-based systems to diagnose cases early. The current work presented fully automatic diagnosis systems that exceed manual techniques to avoid misdiagnosis, reducing time, effort and cost. The proposed system classifies DR images into five stages—no-DR, mild, moderate, severe and proliferative DR—as well as localizing the affected lesions on retain surface. The system comprises two deep learning-based models. The first model (CNN512) used the whole image as an input to the CNN model to classify it into one of the five DR stages. It achieved an accuracy of 88.6% and 84.1% on the DDR and the APTOS Kaggle 2019 public datasets, respectively, compared to the state-of-the-art results. Simultaneously, the second model used an adopted YOLOv3 model to detect and localize the DR lesions, achieving a 0.216 mAP in lesion localization on the DDR dataset, which improves the current state-of-the-art results. Finally, both of the proposed structures, CNN512 and YOLOv3, were fused to classify DR images and localize DR lesions, obtaining an accuracy of 89% with 89% sensitivity, 97.3 specificity and that exceeds the current state-of-the-art results.

Abstract-Task scheduling in data centers is a complex task due to their evolution in size, complexity, and performance. At the same time, customers' requirements have become more sophisticated in terms of execution time and throughput. Against this background, this work presents a new model of resource allocation that optimizes task scheduling using a multi-objective optimization (MOO) and particle swarm optimization (PSO) algorithm. In more detail, we develop a novel multi-objective PSO (MOPSO) algorithm, based on a new ranking strategy. The main insight of this algorithm is that the tasks are scheduled to the virtual machines to minimize waiting time and maximize system throughput. The algorithm leads to a reduction in execution time of 20%, a reduction the waiting time of 30%, and shows improvements of up to 40% in throughput compared to the current state of the art.

In order to increase the reliability, accuracy, and efficiency in the eHealth, Internet of Medical Things is playing a vital role. Current development in telemedicine and the Internet of Things have delivered efficient and low-cost medical devices. The Internet of Medical Things architectures being developed do not completely recognize the potential of Internet of Things. The Internet of Medical Things sensor devices have limited computation power; in case if a patient is using implanted medical devices, it is not easy to recharge or replace the devices immediately. Biosensors are small devices with limited energy if these devices do not wisely utilize the energy may drain sharply and devices become inactive. The current medical solutions place the bulk of data on cloud-based systems that ultimately creates a bottleneck. In this article, an energy-efficient fog-to-cloud Internet of Medical Things architecture is proposed to optimize energy consumption. In the proposed architecture, Bluetooth enabled biosensors are used, because Bluetooth technology is an energy efficient and also helps to enable the sleep and awake modes. The proposed fog-to-cloud Internet of Medical Things works in three different modes periodic, sleep–awake, and continue to optimize the energy consumption. The proposed technique enabled the sensing modes that gathers the patients’ data efficiently based on their health conditions. The sensed data are transmitted to the relevant fog and cloud devices for further processing. The performance of fog-to-cloud Internet of Medical Things is evaluated through simulation; the results are compared with the results of existing techniques in terms of an end-to-end delay, throughput, and energy consumption. It is analyzed that the proposed technique reduces the energy consumption between 30% and 40%.

No abstract