Cyberbully Image and Text Detection using Convolutional Neural Networks

Social media is getting more and more popular in our day to day life. The popularity of social media affects the people involved in it. This makes the technology to do the work or to feel smarter and but only makes us lazy. Therefore, in this robust, discriminative and numerical representation, learning of text messages is a critical issue. Hence, the existing system helps to detect the cyberbully words using Naive Bayes Classifier. The output is classified into cyberbully and not cyberbully words from the Instagram dataset and accuracy is calculated. The proposed framework deployed for detecting negative online interactions in terms of abusive contents carried out through both text and images. This proposed technique is going to detect the cyberbully image and text on the Instagram dataset using Convolutional Neural Network and Bag of words techniques along with the existing technique. Thus, the detected cyberbully words are further classified using Naive Bayes classifier such as Harassing, Insulting, Trolling and Threatening. The combination of text & image analysis techniques is considered an appropriate platform for the detection of potential cyberbully threats.