VideoMudra – Digital Watermarking Solution

The VideoMudra tool can be used to embed a unique watermark on a video file to enable piracy tracking and detection. The uniqueness of the solution lies in the fact that the tool can customise the detection based on the actual attack operated on the particular instance of watermarked video.

VideoMudra: Business Challenges

In most existing watermarking techniques, all the “smarts” in watermarking are applied up-front, in creating and embedding the watermark. However, the robustness of a watermark-embedding algorithm may be compromised in the event of multiple attacks. Consequently, assured detection is not always guaranteed.

The embedding and subsequent template-based detection mechanism used by VideoMudra allows the detection to take into account the actual attack used, or a small range of possible attacks if the exact attack is not known. This significantly raises the chances of an attacked video being accurately identified.

VideoMudra: Technical Details

VideoMudra is a digital video watermarking solution that can successfully trace the source of the piracy even if the video is manipulated to a certain extent. It employs both digital signal processing and computer vision techniques to detect an invisible watermark that has been embedded in the video prior to distribution. VideoMudra uses techniques such as feature transforms and correlation filters for the estimation of attacks and watermark detection.

System Requirements

• Watermark Embedding and Piracy Detection – Intel Quadcore with 4 GB RAM, 512 GB HDD, Linux OS, OpenCV, LAPACK, FFTW, Qt4  libraries

VideoMudra: Benefits

The VideoMudra solution automates the identification of the source of piracy, thereby needing minimal human involvement. It allows customers to protect their copyrighted videos, and thus, their revenue streams. Some key benefits afforded by the VideoMudra solution are listed below.

• Increased detection of accuracy through customised templates
• Real-time watermark embedding
• Invisible watermark that does not affect the video quality
• High tolerance to rotation, scaling, filtering, cropping and compression attacks
• Estimation capability of several attacks, thereby realising faster detection
• Highly secure watermark through customer-dependent watermark generation
• Low false positive rates
• Ease of integration into other frameworks
• Independent of video formats

For more information about VideoMudra, refer resources section.