Skin cancer classification using NASNet

The importance of making an early diagnosis in both the prevention and treatment of skin cancer cannot be overstated. A very effective medical decision support system that can classify skin lesions based on dermoscopic pictures is an essential instrument for determining the prognosis of skin cancer. In spite of the fine-grained variation in the way different types of skin cancer appear, Deep Convolutional Neural Networks (DCNN) have made great strides in recent years toward improving the ability to detect skin cancer types using dermoscopic images. It has been claimed that there are a few different machine learning techniques for the accurate diagnosis of skin cancer using medical photos. A good number of these methods are predicated on convolutional neural networks (CNNs) that have already been trained, which makes it possible to train models using only a small quantity of available training data. However, because there are so few sample images of malignant tumors available, the classification accuracy of these models is still typically severely restricted. The primary purpose of this study is to construct a DCNN-based model that is capable of automatically classifying different types of skin cancer as either melanoma or non-melanoma with a high level of accuracy. We propose an optimized NASNet architecture, in which the NASNet model is enhanced with additional data and an additional basic layer that is employed in CNN is added. The strategy that has been proposed enhances the model's capacity to deal with incomplete and inconsistent data. A dataset of 2637 skin images are used to demonstrate the benefits of the technique that has been proposed. We analyze the performance of the suggested method by looking at its precision, sensitivity, specificity, F1-score, and area under the ROC curve. The accuracy of Optimized NASNet Mobile and NASNet Large provides an accuracy of 85.62% and 83.98%, respectively for Adam optimizer.