Search

Skin cancer occurs due to the abnormal development of the skin cells. It is extremely important to identify this change in skin cells as soon as possible otherwise it is harmful to human life. Among all, malignant melanoma or melanoma is a more dangerous skin cancer. The automatic and accurate identification of melanoma is highly essential as it helps in the diagnosis process. The proposed model uses the Histogram Equalization (HE) and Adaptive gamma correction with weighting distribution (AGCWD) techniques for enhancement of the texture region to obtain better segmentation results. Further, the proposed model focuses to detect the skin lesion automatically by combining DeepLabV3+ with the different base networks such as ResNet 18, ResNet 50 and MobileNetV2. The proposed model is tested using a variety of images from the ISIC 2016, ISIC 2017 and ISIC 2018 datasets. The proposed model is evaluated by comparing with the existing approaches. [ABSTRACT FROM AUTHOR]

Cognitive radio is a promising technology for efficient spectrum utilization. It explores dynamic spectrum access features while satisfying interference constraints. In this work, a joint power and spectrum allocation algorithm is proposed to maximize the cognitive network throughput while satisfying interference constraints of both primary and secondary users in the network. Evolutionary algorithms are used to solve the joint power and spectrum allocation problem. Furthermore, the algorithmic performance is compared in terms of quality of solution. And also we optimized the maximum utilization of the network and capacity of each user simultaneously by using Multi‐Objective Differential Evolution (MODE) and Nondominated Sorting Genetic Algorithm II (NSGA‐II). Simulation results show that the pareto optimal fronts provide the trade‐off solutions between total network utilization and individual sum capacity of user. [ABSTRACT FROM AUTHOR]