Using Transfer Learning to Classify Pornographic Images

Along with the growth of technology and social media, need to control and isolate people, especially children, from pornography has become quintessential. The existing Convolutional Neural Networks provide ways to distinguish between adult and non-adult images, but they are far too complex to use them. We use Transfer Learning as a simple approach to prove that we can easily classify adult images using simple models rather than existing complex image classification models or CNNs. The ease is that we will not build a complete architecture from scratch, rather, we will apply Transfer Learning to make use of complex architectures of existing CNNs.