Advertisement embedding for online videos

• Main Author: Han, Wei.
• Other Authors: Yuan Junsong
• Format: Final Year Project
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/54506
• Institution: Nanyang Technological University

Nowadays, most existing online video players and websites provide search function. However, the search function can only be realized by key word matching. Often, a user want to search for videos that contain a certain object which is hard to describe in key words. In this situation, a search method using a query image as a key is needed. The method to find and locate a target object in a large collection of images or videos is called Visual Object Search. Visual Object Search can also be used for embedded online videos advertisement. It enables online video websites to efficiently and accurately embed advertisements of a commercial product to corresponding video frames that contain the commercial product. In this project, a video search algorithm was developed based on a Visual Object Search method using Randomized Visual Phrase via Random Partition, which is proposed and developed by a research team lead by Professor Yuan Junsong from NTU. A Windows 8 Store App was built implementing the video search algorithm. The App enables users to search and locate a query image in real time in a large collection of videos. The search result of the App is accurate enough for practical application. In addition, a video advertisement interface was also designed and developed in this project.