Hate Speech Detection using Convolutional Neural Network Algorithm Based on Image

Hate speech is words behavior that can cause an attitude of violence and anarchy against other individuals or groups. The internet has become necessary in this day and age, so internet morals need to be considered. However, several parties deviate from using the internet to spread hate speech, such about race, ethnicity, and religion. Nowadays, hate speech detection systems are usually through text but hate speech detection through images tends to be rare. For that reason, this study is aimed to detect whether there is hate speech or not in the selected image. This project uses the Convolutional Neural Network (CNN) algorithm and Deep Learning method to classify the aspect of hate speech contained in an image and recognize any hate speech on the image through the existing text. After this application is developed, the machine learning system can detect some hate speech on an image that contains the text. It achieves about 95.89% accuracy and 94.43% precision. After that, the authors hoped that the authorities could reduce hate speech in the community and follow up more quickly.