International Journal of Computer Applications (0975 – 8887) Volume 184 – No. 42, January 2023 5 Skin Cancer Classification using VGG-16 and Googlenet CNN Models Suryakanth B. Ummapure, PhD Department of Computer Science Govt. First Grade College Shahapur, Karnataka, India Ravindrakumar Tilekar Department of Computer Science Govt. First Grade College Bidar, Karnataka, India Satishkumar Mallappa, PhD Sri Satya Sai University for Human Excellence Kalaburagi Campus, Navanihal , Kamalapur Karnataka ABSTRACT Skin cancer with a high fatality rate is called melanoma. Due to the great degree of similarities among the many forms of skin lesions, a proper diagnosis cannot be made. Dermatologists can treat patients and save their lives by accurately classifying skin lesions in their early stages. This paper proposes a model for highly accurate skin lesion classification. The proposed model made use of transfer learning models known as GoogleNet and vgg16. This model efficiently distinguished between benign and malignant cancerous skin lesions, those are the two distinct classes of skin diseases. The 1800 benign cancer images and 1498 malignant cancer images that were retrieved from the internet were taken into account for this proposed strategy. The VGG16 has obtained the highest recognition accuracy in the result accessing, with recognition rates of 99.62% for training and 84.97% for validation. Keywords Googlenet, VGG16, Skin Cancer, Deep Learning 1. INTRODUCTION Skin cancer [1] [2] is a form of cancer that arises when DNA damage that has not been repaired causes abnormal growth in skin cells. It is best to identify skin cancer early since it is more manageable in its initial stages, despite the fact that it spreads progressively to other regions of the body. Early detection of skin cancer signs is necessary due to the increased rate of the disease, high death rate, and high expense of medical care. The two types of skin cancer are classified as malignant and benign. Traditional computer vision methods are frequently used as a classifier to extract numerous features, such as shape, size, color, and texture, in order to identify cancer. Due to the availability of deep learning techniques, these features are now not very frequently utilized. Feature classification and feature detection are two most common deep learning methods. This method is capable of extraction the potential features of its own and those features are called CNN features. Hence, from skin cancer images, CNN features were extracted and efficiently utilized for the classification purpose. The flow of this paper is in this manner section-2 describes earlier works carried out in the relevant area. The experimental details are given in section-3. The outcome of the experimental results and discussions are presented in section-4. At last, section 5 ends the paper with conclusion and future scope of this work. 2. PREVIOUS WORK If skin cancer is found in its early stages, treatment may be given. The patient's skin observations can aid surgeons in making the best treatment options based on skin cancer images. Numerous studies on the identification and classification of skin cancer have been published in the literature. The paper published by Shahin et.al[3], have proposed the deep learning based skin cancer classification they have used HAM10000 dataset and obtained the 96.16% recognition accuracy for training and for testing they have achieved the 91.96% recognition accuracy respectively and they have compared the model with other models such as Resnet, AlexNet, VGG-16,MobileNet DenseNet, etc. In Jinnai et.al.[4], they have reported the work on skin cancer development for the pigmented skin lesions by applying the deep learning method from this they have got an highest of 91.5% highest recognition accuracy. The integrated design of deep features fusion based skin cancer classification is reported by Amin et al.[5], they have used the deep features and applied PCA on the features and obtained the 99.00% recognition accuracy. An seven ways skin cancer classification by using Mobilenet is reported by Chaturvedi et al.[6], by using pretrained model named as Mobilenet on 2014 ImageNet Challenge dataset and obtained the overall accuracy of 83.10%. A review paper on skin cancer classification by using various methods; deep learning, and CNN is given by Manne et al.[7]. The interpreted deep learning method to segment and classification of non-melanoma skin cancer is reported by Thomas et al.[8], from this method they have obtained the 97.9% recognition accuracy. Filali et al.[9] is shown the best use of hand crafted and CNN features for skin cancer classification and they have used the PH2 dataset and obtained the 98% recognition accuracy. A novel approach is reported in the work of Singh et al.[10], they have reported the transfer learning (TL) framework Transfer Constituent Support Vector Machine (TrCSVM) from this method they have obtained the 98.82% overall recognition accuracy. 3. SIMULATION DETAILS In the process of carrying out this proposed experiment the two popular pretrained CNN models; Googlenet and VGG16 applied on 3298 skin cancer images. The pretrained models are widely used to carry out the image classification kinds of work. To obtain the maximum recognition accuracy these models may be used. These are the models which are built by someone else and others were used to test their data on them with the intention of getting maximum recognition accuracy to fulfill the research objective belongs to many research areas. By using these pretrained models for others experiment is sometime known as transfer learning also.