International Science and Technology Journal

In recent decades, melanoma, a lethal form of skin cancer, has become more common internationally. Due to the recovery strategy used in the medical field, dermoscopic image-based automatic skin lesion detection remains a challenging and complicated task. This difficulty in diagnosing lesions can be attributed to various factors, such as the lesions revealing diverse characteristics, including uncertain borders, inadequate color assessment, shape variations, positional dependence, and complex structures. To prevent the growing public health burden from spreading to other body organs and potentially save many lives, early detection and appropriate treatment are essential among medical professionals and researchers. An individual may develop melanoma if there is an abnormal change in the appearance of the skin. To achieve more effective cancer detection, dermatology expertise should be integrated with computer vision strategies. Consequently, it is crucial to create a variety of detection strategies to aid medical experts in diagnosing early-stage cancer. This paper provides a thorough, methodical analysis of machine learning techniques for early skin cancer detection, examining published studies. It offers an overview of an artificial intelligence-assisted evaluation approach for diagnosing skin cancer. To address existing challenges in skin cancer detection, this paper proposes a model integrating improved techniques across preprocessing, feature extraction, and classification. The proposed model utilizes adaptive histogram equalization to enhance image clarity, combined with feature extraction methods such as the ABCD rule, GLCM, and HOG, alongside deep learning-based autoencoders. For classification, it employs an ensemble approach that integrates CNNs and SVMs, leveraging transfer learning from pre-trained models and implementing a multi-class framework. This model aims to enhance detection accuracy, reduce costs, and improve usability. The review aims to provide researchers with the latest advancements in machine learning for cancer diagnostics, offering a comprehensive understanding of current and emerging techniques......................... Keywords:........... Skin Cancer, Melanoma Detection, Artificial Intelligence, Machine Learning, Deep Learning ،Neural Network Melanin.