Abstract:
Social Media has become a tool defining this era, daily the means to connect with each other are expanding from different paradigms. YouTube holding a position of prominent video sharing, allows users to connect and share anything they are willing to share. This excessive actions taken by YouTube onlookers, such as commenting and liking, triggers certain motives for video up loaders to instil motives to gain profit or fame. Even though such motives are encouraged, apart from tools aiding video creation, no tool has been created to support videographers in predicting their video’s outcome once uploaded. This project incorporates deep learning to video domain. The Anticipator allows a video uploaded to be analysed to recognize it attributes further to retrieve the comments that are present in YouTube related to the images recognized. The application is built on a CNN which is the most effective neural network for classifying the images, with the integrating of the YouTube API as well as a database built on firebase and the frontend designed with Flask. This system was tested and evaluated using standardized approaches which are qualitative and quantitative methods. The ultimate Anticipator system performs all functions completely and as expected efficiently.
