Aquila coyote-tuned deep convolutional neural network for the classification of bare skinned images in websites

Pornographic websites become increasingly obdurate via disguising and misleading the under-aged people, which obstruct the development of a healthy and safe network environment. The conventional strategies of bare skin classification perform the detection process through a single feature of these sites, which would be unsuccessful to solve the more complicated and tedious situations. Hence, this research focuses on proposing a well-adapted port image classification model using an optimized deep convolutional neural network (Deep CNN). The significance of this research relies on the proposed Aquila Coyote (AqCO) optimization algorithm, developed with the integration of Aquila hunters and the coyote hunters, which engages in tuning the hyper-parameters of the Deep CNN classifier. Moreover, the significant features of the image are utilized by the classifier, which further boosts the classification accuracy of the proposed model. The analysis is done based on performance parameters, such as accuracy, sensitivity, and specificity in such a way to evaluate the efficiency of the proposed porn image classification model. The maximal accuracy of the proposed model is 95.91% for the B-Praneeth dataset, which is high as compared to the existing methods of porn image classification methods.