Multi-Input Melanoma Classification Using Mobilenet-V3-Large Architecture

All over the globe there exists a serious problem with skin cancer, most especially melanoma; a malignancy which is known to behave aggressively and can metastasize. Detecting this early is the key to saving lives. This study introduces a new method of classifying melanoma using an advanced model known as MobileNet-V3-Large. This technique differs from others in that it considers both the images of the skin lesion and tabular data including factors consisting of the patient’s approximate age, gender and the location of the lesion on the body of the patient. Such an approach empowers the predictions of whether the skin lesion may be malignant or benign. Tested on a huge collection consisting of skin images combined with tabular data, it was established that this method outperforms others already existing. The results of this study showed that a high accuracy of 99.56% was achieved using the proposed model. This study indicates that utilizing the multi-input method will substantially enhance diagnosis for melanoma hence reducing mortalities in the future.