The selection of optimal relay node ever remains a stern challenge for underwater routing. Due to a rigid and uncouth underwater environment, the acoustic channel faces inevitable masses that tarnish the transmission cycle. None of the protocols can cover all routing issues; therefore, designing underwater routing protocol demands a cognitive coverage that cannot be accomplished without meticulous research. An angle-based shrewd technique is being adopted to improve the data packet delivery, as well as revitalize the network lifespan. From source to destination, one complete cycle comprises three phases indeed; in the first phase, the eligibility of data packet belonging to the same transmission zone is litigated by Forwarder Hop Angle (FHA) and Counterpart Hop Angle (CHA). If FHA value is equal or greater than CHA, it presages that the generated packet belongs to the same transmission zone; otherwise, it portends that packet is maverick from other sectors. The second phase picks out the best relay node by computing a three-state link quality with prefix values using the Additive-Rise and Additive-Fall method. Finally, the third phase renders a decisive solution regarding exorbitant overhead fistula; a packet holding time is contemplated to prevent the packet loss probability. Simulation results using NS2 have been analyzed, regarding packet delivery ratio, packet error rate, communication overhead, and end-to-end delay. Comparing to HHVBF and GEDAR, USPF indeed has outperformed, leading into the evidence of applicability’s favor.

Retinal blood vessels, the diagnostic bio‐marker of ophthalmologic and diabetic retinopathy, utilise thick and thin vessels for diagnostic and monitoring purposes. The existing deep learning methods attempt to segment the retinal vessels using a unified loss function. However, a difference in spatial features of thick and thin vessels and a biased distribution creates an imbalanced thickness, rendering the unified loss function to be useful only for thick vessels. To address this challenge, a patch‐based generative adversarial network‐based technique is proposed which iteratively learns both thick and thin vessels in fundoscopic images. It introduces an additional loss function that allows the generator network to learn thin and thick vessels, while the discriminator network assists in segmenting out both vessels as a combined objective function. Compared with state‐of‐the‐art techniques, the proposed model demonstrates the enhanced accuracy, sensitivity, specificity, and area under the receiver operating characteristic curves on STARE, DRIVE, and CHASEDB1 datasets.

The accomplishment of sustainable communication among source and destination sink node is a rigors challenge and even establishing bodacious communication link between these nodes is nothing short of a miracle because data routes are governed by the underwater environment. Energy consumption has a significant influence as all active devices rely on the battery. As cost-effective data packet transmission is established as a norm, no charging or replacement can be achieved. Hop link evaluation and shrewd connection discovery by way of a resurrecting linking element were just a genuinely grim task, and only feasible to create the extra powered energy pods (URR-SAEP) that had never been carried out before after detailed study. After packet transfer, the sensor node performs the link inspection process, and when a link is deemed shaky at less than or equivalent to 50 percent of capacity, the target node incorporates its residual capacity status and returns it to the source node that attaches other unoptimizable energy pods to improve only the targeted node link from 50 percent to 90 percent. Performance evaluation using NS2 with Aqua-Sim 2.0 simulator has been obtained comparing with DBR and EEDBR protocols in terms of point-to-point delay, Packet dissemination ratio, Network lifespan and Energy Diminution.

Remote sensing image scene classification has drawn significant attention for its potential applications in the economy and livelihoods. Unlike the traditional handcrafted features, the convolutional neural networks (CNNs) provides an excellent avenue in obtaining powerful discriminative features. Although tremendous efforts have been made so far in this domain, there are still many open challenges in scene classification due to the scene complexity with higher within-class diversity and between-class similarity. To solve the above-mentioned problems, D-CapsNet is proposed to learn the richer and more robust features for scene classification. It is an end to end network with four types of layers and incorporates visual attention mechanisms. Its diverse capsules encode different properties of complex image scenes, including deep high-level features, spatial attention based on the fusion of multilayers features, both spatial and channel attention based on high-level features, and their fusion. Experiments on three image scene datasets demonstrate that D-CapsNet outperforms other baselines and state-of-the-art methods with a significant improvement in both classification accuracy and speed.

No abstract