Cyberbullying Image Classification using Transfer Learning Model

Abstract

Abstract: The impact of cyberbullying is immeasurable on the lives of victims as it is very subjective to how the person would tackle this. The message may be a bully for victims, but it may be normal for others. The ambiguities in cyberbullying messages create a big challenge to find the bully content. Some research has been reported to address this issue with textual posts. However, image-based cyberbullying detection has received less attention. This Project aims to develop a model that helps to prevent image- based cyberbullying issues on social platform posts. We proposed a transfer learning-based automated model to detect image-based cyberbullying posts from the social platform. The transfer learning models are capable of extracting hidden contextual features from cyberbullying posts. Our experiment consists of two sets of datasets (i.e.) images consisting of cyberbullying and non cyberbullying images. The datasets can be useful for future researchers to extend the research. Finding the best-suited model to detect the bully images is a challenging task, hence experimented with both DL and transfer learning models to find the best model. The experimental outcomes confirmed that the transfer learning models are the better choice for predicting image-based cyberbullying posts.