More and more genetic and metabolic abnormalities are now known to cause cancer, which is typically fatal. Any particular body part may become infected by cancerous cells, which can be fatal. One of the most prevalent types of cancer is skin cancer, which is spreading worldwide.The primary subtypes of skin cancer are squamous and basal cell carcinomas, as well as melanoma, which is clinically aggressive and accounts for the majority of fatalities. Screening for skin cancer is so crucial.Deep Learning is one of the best options to quickly and precisely diagnose skin cancer (DL).This study used the Convolution Neural Network (CNN) deep learning technique to distinguish between the two primary types of cancers, malignant and benign, using the ISIC2018 dataset. The 3533 skin lesions in this dataset range from benign to malignant, and nonmelanocytic to melanocytic malignancies. The images were initially enhanced and edited using ESRGAN. The preprocessing stage involved resizing, normalising, and augmenting the images. By combining the results of numerous repetitions, the CNN approach might be used to categorise images of skin lesions. Several transfer learning models, such as Resnet50, InceptionV3, and Inception Resnet, were then used for fine-tuning. The uniqueness and contribution of this study are the preprocessing stages using ESRGAN and the testing of various models (including the intended CNN, Resnet50, InceptionV3, and Inception Resnet). Results from the model we developed matched those from the pretrained model exactly. The efficiency of the suggested strategy was proved by simulations using the ISIC 2018 skin lesion dataset. In terms of accuracy, the CNN model performed better than the Resnet50 (83.7%), InceptionV3 (85.8%), and Inception Resnet (84%) models.

Most industries are now switching from traditional modes to cloud environments and cloud-based services. It is essential to create a secure environment for the cloud space in order to provide consumers with a safe and protected environment for cloud-based transactions. Here, we discuss the suggested approaches for creating a reliable and safe environment for a surveillance cloud. When assessing the security of vital locations, surveillance data is crucial. We are implementing machine learning methods to improve cloud security to more precisely classify image pixels, we make use of Support Vector Machines (SVM) and Fuzzy C-means Clustering (FCM). We also extend the conventional two-tiered design by adding a third level, the CloudSec module, to lower the risk of potential disclosure of surveillance data.In our work we evaluates how well our proposed model (FCM-SVM) performed against contemporary models like ANN, KNN, SVD, and Naive Bayes. Comparing our model to other cutting-edge models, we found that it performed better, with an average accuracy of 94.4%.