Design team members: Brian Cheung, Matthew Jama, Simon Huang
Supervisors: Dr. David Clausi
Background
There are a variety of different streaming video services on the internet. These services host a range of video formats, such as amusing and very short clips, medium length films and full feature-length films. Currently, there is not an existing product that allows online video users to easily bookmark specific parts of a video, share the bookmarks with friends or return to the bookmarked clips at a later time. There have been some attempts to index and bookmark parts of video, but no easy to use and complete integration currently exists.
Online streaming video services such as Youtube, Vimeo and Metacafe allow users to upload personal videos. An existing method of bookmarking is to cut the video up into independent sections and upload the video in smaller parts that can be easily linked and shared. Each of the parts can be labeled with an appropriate tag or category. However, this approach lacks continuity, and users will be unable to watch the full video without jumping from one video segment to another. Furthermore, this method does not provide the user with any context regarding the full video and how all the parts fit together.
If a method could be developed that would allow users to easily bookmark, share, and retrieve video content, an individual watching a movie would be able to tag significant moments like famous scenes or funny quotes so that they can conveniently re-watch the scenes in the future or share scenes with his or her friends. A parent recording his or her child’s sports event, school play, or other significant event would be able to upload the camcorder footage to the internet and share it with friends/family, with video highlights bookmarked so that the viewer may skip to the most interesting parts of the video. A busy executive could watch the all the highlights with his child from the school's basketball game on his phone while waiting in the airport lounge. An Internet user who needs specific video footage and would like to find all video clips pertaining to a specific theme or tag. For example, a student researching American history could search for video clips containing presidential speeches and quotes.
Project description
Skipbird is an online portal that allows users to search for videos from locally hosted media and a variety of online websites. Once a video is found, users can watch the video, see other people’s shared bookmarks or bookmark parts of videos for themselves.
Final UI Design
Users can load videos from youtube, vimeo, vevo, or locally-hosted videos. Bookmarking will be crowd-sourced and voting algorithms will be used to populate a list of bookmarks that are appropriate for a video.
Design methodology
Skipbird is built to industry standards, using as many open source packages from ground up. A proof-of-concept is hosted on three servers, with one server hosting the web front-end, one server handling the databases and one final server hosting uploaded media.
The webpage is built on HTML5 and CSS and is designed for maximum compatibility with all modern browsers. The player is intended to be HTML5 compatible, with fallback modes to an open-source flash-based player.
All functions on the webpage are designed to be performed asynchronously, to minimize server load and to maximize scalability. The bookmarking is performed using PHP, which communicates with the business logic layer on the web server, which communicates with the database access layer on the database server, which pulls the information from the database.
When a new video is loaded, the player asynchronously loads the video from the online website or locally-hosted source, while the webpage sends an AJAX request to asynchronously load the bookmarks. Once the bookmarks are returned, they are parsed and refreshed within the left window.
Proof of Concept