ENCVIDC: an innovative approach for encoded video content classification.

NETFLIX Inc.; STREAMING video & television; RANDOM forest algorithms; LEAKS (Disclosure of information); VIDEOS; CLASSIFICATION; VIDEO compression

With the increase in the sum of online video viewers on the internet every day, the video service providers are getting interested to know about the nature of the content being viewed through the supplied network in order to accomplish their business associated objectives that may include the user's internet behavior profile, etc. Due to the widespread use of encoded video streaming techniques, the network video traffic classification has turned out to be a challenging task. As devoid of the authentic decryption key, it is impossible to comprehend the actual content viewed by the user. However, the current advances in machine learning have demonstrated the fact that encryption can also lead to certain information leak which yields promising results in determining the actual transmitted content between the two communicating parties. This research proposes a classifier for determining the encrypted video content over different streaming sites such as YouTube, Netflix and Dailymotion. We demonstrated that an eavesdropper can determine the stream video content even if the traffic is encrypted by identifiable patterns extracted from the captured traffic. We used different machine algorithms for the task and conducted a series of tests, demonstrating that our classification based on Random Forest showed accuracy greater than 98% and has the ability to execute all the network-related business objectives of any enterprise network. [ABSTRACT FROM AUTHOR]

Copyright of Neural Computing & Applications is the property of Springer Nature and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)