Managing Spam Images on Android: An Approach Utilizing Machine Learning and NLP

In today’s digital communication environment, images have become an integral part of our interactions with others. However, a significant portion of these images have no practical value and are mainly spam images, usually greeting images like ‘good morning’ and ‘good night’, etc. This spam image originates from popular social media applications like WhatsApp, which have flooded users’ mobile devices with unwanted spam images and also consumed storage space but also degraded the performance of the Android/Mobile device. This research paper focuses on the implementation of spam image detection in the Android ecosystem. This application works completely offline, thus reducing internet dependency. The suggested approach incorporates methodologies like natural language processing, optical character recognition and machine learning. Spam detection in Marathi, Hindi, English languages are supported and can be further increased to other languages as well. The NLP model in this application is trained on our own dataset; this creates flexibility for supporting other languages.