"If you want to discover the true character of a person, you have only to observe what they are passionate about". Automatic evaluation of personality traits from the posts posted on various social media platforms is an interesting field of research. The interests of a person are reflected in the content of images which users post to their social media accounts. Popular personality prediction systems classify the personality of a person on the basis of five broad personality traits Openness, Conscientiousness, Extraversion, Agreeableness and neuroticism (also known as the Big Five Personality Traits). This paper proposes a novel approach of predicting a person's personality reflected by his/her interests. The project detects significant features in each image posted by the user in his/her social media profile and then maps the features detected in the image to a predefined class which is based on the frequent interest domains normally observed in social media. For example, a person inclined towards adventure might post a higher number of photos related to adventurous activities like trekking, surfing, etc. The classes, for instance, can be; animal lover, foodie, adventurous etc. The significant features in an image posted by, for example, an animal lover might be a dog, a cat, or any other animal. Normally, a person is more likely to make friends with other person if they share similar interests. This project helps to connect an individual to other people sharing similar interests with him/her and make social media more "familiar" to the user.
