Parags Homepage

3D contents such as 3D models and 3D motion capture data are being readily used in the entertainment and medical industry. In the entertainment industry they are being used to develop animations, video games, movies and commercials. In medical sector they are used for applications such as 3D Images, biomechanical research. Since these data sets involve investment in terms of money, time and human effort preserving them is important. Loss due to theft can result in their piracy or their tampering can result in loss of information important to medical science or patients being cured. Therefore, it is imminent that we need to content protect the same.

Illegal usage deals with ownership of the media wherein copyright of the media is infringed in case someone steals the data is uses it without the permission of the owner. This can occur incases where data is stolen (video games) and being sold as pirated copies. So one has to come with a system where it can be prove in the court of law one needs to prove the ownership of the content.

Tampering involves changing the meaning of the multimedia. This can occur in cases where meaning of the data is changed and since the data is valued for its meaning the content is lost. For example, in case of 3D data such as tooth model data can change in meaning due to tampering or machine resulting in accuracy of medical record. To capture this tampering one needs a method to recover from the loss. In addition, incase errors are discovered, we need methods to reconstruct the data to regain the loss.

Problem With Secure Server ? One of the methods is to secure the digital media on a secure server whose access is limited. Users can securely login into the secure server with a predefined cryptographic method such as SSL (secure socket layer) and gain access to the digital media. However, once the media is at the user’s computer one cannot guarantee that it will not be subject to illegal usage, since the user can practice piracy or his/her computer can get compromised to further problems. A similar scenario can occur in medical records, where data being stored in the database gets corrupted due a machine error or someone hacks in the machine and tampers the records. In both the cases, we see that keeping the records on a secure computer does not always guarantee security. Therefore, we need additional mechanisms which add an extra layer of security to content protect the media.

Digital watermarking is one of the methods that is used to content protect digital media. The idea is to encode patterns ‘watermarks’ inside data without losing the meaning of the media. The presence of the watermark uniquely identifies the media as original, which can help in proving as an evidence for its uniqueness in copyright issues. Also, in case the meaning of the data set is tampered with, the loss the pattern watermark will help as an evidence for tampering. Since it is possible to tamper proof the data set we can authenticate the meaning of the digital media. Another method to authenticate the data set is to generate signatures which are patterns from the data set. In case, we have to check if the data is original, we generate a signature and compare it with a previous generated signal. The similarity between the signatures authenticates the originality of the digital media. This process of generation of signature is termed as Finger-printing.

Why the choice of solving problems related to 3D data is justified? Different methods have been developed to content protect digital media with main focus on images, text, video and audio. However, little progress has been made for 3D content comprising of 3D models and 3D time series data. Although different watermarking and finger printing techniques have been developed for other kinds of media, they cannot be directly applied to 3D content. One of the reasons is dimensionality since other kinds of media are not generalized to handle problems related to higher dimensional data. Secondly, since the solutions for 3D content can be developed for n-dimensional data, for example n-dimensional time-series solutions is applicable to 1D audio data and 1D medical signal such as Electrocardiogram. So with factors such as dimensionality and applicability of the solution we find developing methods for 3D data more challenging and promising. However, another important purpose of developing techniques for this kind of data is their applicability. We now define these data set and related applications and show the reasons for protecting them.