Abstract
Multimedia forensics can be defined as the science that tries, by only analysing a particular digital asset, to give an assessment on such a content and to extract information that can be useful to address and support an investigation linked to the scene represented in that specific digital document. The basic idea behind multimedia forensics relies on the observation that both the acquisition process and any post-processing operation leave a distinctive imprint on the data, as a sort of digital fingerprint. The analysis of such a fingerprint may permit to determine image/video origin and to establish digital content authenticity.
TopMultimedia forensics can be defined as the science that tries, by analysing a digital asset, to give an assessment on such a content and to extract information that can be useful to address and support an investigation linked to the scene represented in that specific digital document. Multimedia forensics has to be able to develop efficient instruments to deal with the disparate digital devices that can generate images and, above all, with the different processing tools that allows also an unskilled user to manipulate digital goods. Hereafter two basic approaches are introduced, then the various kinds of data that multimedia forensic tools could have to face with are presented. After that, some possible application scenarios where these technologies could be claim to operate are described and finally a wide look to which are the possible digital fingerprints to be searched for in a multimedia content is given.
Key Terms in this Chapter
Data Authenticity: Digital data can be assumed to be authentic if it is provable that it has not been corrupted after its creation. In a strong sense, any processing means corruption, that is digital data to be authentic must be only the outcome of an acquisition process of a real world scene without any successively processing; but in a wide sense, authentic data must accordingly represent a real world scene and even if some processing has been probably applied the meaning of the scene must not be modified. Data authenticity also means that a digital object is indeed what it claims to be or what it is claimed to be.
Source Identification: Given a digital asset, it is possible to trace the device that has produced the data. In particular, by focusing on visual data, source identification refers to the recovery of the type of used imaging devices between digital cameras, scanners, mobiles, computer graphic technologies, or the specific model or brand of such devices
Pattern Noise: A reference pattern noise is a particular digital fingerprint left over a digital image during acquisition. Such pattern is due to the manufacturing process and can be extracted from the images using a denoising filter.
Digital Fingerprints: Any digital asset is characterized by inherent patterns specific of its life history, such patterns, referred as fingerprints, come from the acquisition device producing the data and/or the possible processing suffered by the data.
Multimedia Forensic: Multimedia forensic can be defined as the science that tries, by only analyzing a particular digital asset, to give an assessment on such a content and to extract information that can be useful to address and support an investigation linked to the scene represented in that specific digital document.
Digital Evidences: During a trial a set of evidences are considered before returning a verdict; alongside of witnesses, assertions, and concrete objects, nowadays digital data representing the acquisition and the storage of all the information belonging to the crime scene has to be considered as digital evidences.
Tampering: A tampering operation can be defined as a particular subset of image processing, voluntarily applied, aiming at counterfeiting the meaning of the tampered data or at least at getting something to appear different from what it is really.