The expense of faster and more precise diagnostic skin conditions is still prohibitive and high. Therefore, image processing methods aid in the beginning development of an automated dermatological screening system. The classification of skin disorders relies heavily on the extraction of characteristics. In several methods, computer vision plays a part in the identification of skin conditions. Skin infections are prevalent in Saudi Arabia as a result of the deserts and the hot climate. This paper contributes to the understanding of diseases of the skin detection. Based on image processing, we proposed a method for diagnosing skin problems. Using image analysis, a digital image of the diseased skin region is used to diagnose the type of illness. The only expensive pieces of equipment required for our simple, rapid procedure are a camera and a computer. The approach is based on the inputs of a color image. Resize the image after that using a convolution neural network that has been trained to extract features. Utilizing Multiclass SVM, the feature was subsequently categorized. On dermatological macro-images, a modified sigmoid transform based on EfficientNet may be used to improve the contrast between lesion and background areas. The border between the lesion and background areas of the pixel values may be fixed using the modified sigmoid transform. Aids in increasing the accuracy of skin lesion segmentation. Application in practice as a preliminary processing stage in automated tools for skin cancer detection from dermatological macro-images. On dermatological macro-images, skin lesions may not initially contrast or vary in intensity enough from the surrounding tissue. The contrast is further diminished by the improper exposure of the light at the time the photograph was taken. Segmentation is negatively impacted by low contrast between the lesion and background areas.