The elementary symptom of skin cancers primarily appears as a spot, scaly area on the skin that changes colour, size or shape over a period of time. The detection of skin pigmentation and growth is an early sign of visual cues to identify any indication towards skin cancer. Of all cancers, melanoma is arguably one of the most straightforward to detect early because most melanomas present on visible areas of the skin and mucosa. This pigmentation may be caused for various other reasons which may not lead to carcinogenic structural growth. In many cases, these pigmentations are misunderstood or lately identified as any kind of harmful cell division. The ambiguous visual structure of skin moles may create false positives or negatives to come to a conclusion for identifying the exact disease being it to be benign or malignant. Manual identification of such microscopic slide image is a time and effort constraint activity. Automated AI (Artificial Intelligence)-based computer-aided diagnostics have proved to be reliable in many such cases making healthcare facilities more accessible. In this paper, the research gaps have been addressed and based on these, a new optimum CNN classifier OPT-MobileNet (Optimum MobileNet) has been proposed. A range of state-of-the-art NN (Neural Network) models are compared with the proposed technique for verifying the efficiency and reliability in perfect carcinogenic structure identification. The performance of the proposed technique has been illustrated numerically and graphically for related reference. During the experiment, the proposed methodology has confirmed to identify the skin mole type with in an optimum amount of time and classify the type of skin mole to its exact class. This proposed method of skin mole analysis can serve as a state of Decision Support System (DSS) in medical pathology.