Detection of Falsary Happening on Social Media Using Image Processing: Feature Extraction and Matching

Abstract

In last decade and half, especially generation Z (Gen Z) has witnessed a drastic change in their communication pattern due to paradigm shift in digital media. Social media is the largest medium to connect any part of the world via internet is considered to be the trend changer. Increasing use of mobile internet and easy availability of mobile compatible softwares have changed the way of expression on social media in both apt and inept manner. Forged images known as memes are, nowadays, trending on social media like Facebook, Instagram, Twitter, Whatsapp and what not. These memes on social media are used to appreciate or to troll on particular cause to someone or group of people. The plethora of memes usage leads to image processing where feature extraction and feature matching can be done. Matching of features that are invariant to transformation includes geometric invariance and photometric invariance. This paper presents the comparative analysis of images (memes) matching algorithms of feature, which will definitely helpful for future researchers to identify the best matching technique to be fit in desired area. Authors have also included the implemented results of various traditional matching techniques invariant to translation, rotation, scaling, brightness and exposure.