Cyberbullying Image Classification using Artificial Intelligence for Safer Online Platform

The toll that cyberbullying takes on victims' life is incalculable because a person's approach to it is highly individualized. For some victims, the message might be harsh, but for others, it might be normal. It is quite difficult to find reliable content because of the ambiguities in cyberbullying communications. Several research works have reportedly used a text- content-based methodology to answer this question. However, picture-based cyberbullying detection is now receiving less attention. Enhancing a methodology that permits the prevention of social media posts on image-based cyberbullying issues is the aim of this job. This study provides support for an automated version that searches a malicious social media network for a cyber-symbol image, relying just on switch mastery. By applying fashion knowledge, it is feasible to infer hidden contextual functions from texts containing cyberbullying. Our test includes two photo datasets: pictures of cyberbullying and pictures of non-bullying. The datasets might be helpful to Destiny researchers in expanding their research. It's challenging to find a pleasant and suitable version to recognize bullying photographs, therefore experiment with both transfer the experiment's findings showed that switch learning models are an excellent method for predicting conversations that are only based on images and involve cyberbullying.