Watermarks have been proposed for Copyright Protection of digital images, audio and video and, extensively, multimedia products. They are digital signals that are embedded into other digital signals (carriers). The embedding procedure is called watermarking and the resulting carrier signal is not affected strongly by such an embedding procedure (i.e. watermarks are invisible).

A watermark should represent exclusively the copyright owner of the product and can be detected only by him/her and must be robust to any product modification that does not degrade its quality. Resistance against any intentional attack is required.

Our Methods

Watermarking methods have been developed to sign digital still images, (color or gray scale), digital video, and digital audio signals. They are characterized by the following technical characteristics:

Perceptual Invisibility: The watermark signal is embedded in the digital data producing alterations. These alterations should not degrade the perceived quality. The owner can decide about the strength of alterations. Large alterations are robust and can be detected with great certainty, but they produce degradation of the product.

Complexity: Watermark signals are characterized by great complexity. This is necessary in order to avoid the construction of similar watermarks. Subsequently, watermark determination by a third person , becomes more difficult. Another advantage of "complex'' (complicated) watermarks is that they provide reliable statistical properties and their detection can be shown with great certainty.

Watermark key: Any watermark signal is associated (one by one) with an integer number (or a set of integer numbers) which is the watermark key. This key is used to produce, embed and detect a watermar k. The key is private and characterises exclusively the legal owner of the digital product. The number of available keys is enormous.

Statistical efficiency: The detection of a particular watermark is successful when the suitable key is used. We note that each watermark corresponds to a unique key.

Statistical invisibility: The possession of a great number of digital products, watermarked by the same key, does not dispose the watermark. Different products watermarked by the same key, transfer di fferent watermark signals. We can claim that the extraction of the owner's key by a third person is impossible. Counterfeit watermark keys can not be determined, i.e. keys, which detect a watermark that had never been embedded in the digital image before.

Multiple Watermarking: A sufficient number of different watermarks can be embedded in the same image. Each of these watermarks could be detected by using the corresponding unique key. A creator can use the same key f or watermarking a large set of products.

Robustness: Various kinds of modifications should be performed on a digital product in order to improve their quality or to compress their size or to remove the watermarks. Watermarks use efficient ways to produce invisible and, simultaneously, robust alterations to various modifications like filtering and JPEG (or MPEG) compression. Therefore the watermark is still detected in a modified image that demonstrate high quality.

We have proposed the following watermark types:

• Pseudo-random noisy binary patterns: They are applied to still images. The watermark embedding and detection algorithms are very fast; however, such watermarks do not resist to high JPEG compression ratio and low-pass filtering.
• Low-pass pseudo-random or chaotic signals: They are applied to still images and audio data. Their robustness is satisfactory under filtering and JPEG compression.
• Mixing watermarks: These watermarks represent binary copyright logos, which are embedded in still images and video sequences. The watermark logo can be either reconstructed from the watermarked image or they can be detected statistically. Statistical detection provides certain decisions about the watermark existence even after high JPEG or MPEG compression.
• DCT constraint: The watermark is expressed by constraint in DCT space.
• Region-based watermarks: A pseudo-random watermark is embedded in several regions of the image, and is adapted according to certain features of these regions. It is therefore robust to certain geometric operations like rotation scaling and translation, apart from the usual filtering and compression..
• Circularly symmetric watermarks: A pseudo-random watermark is embedded in a certain band of DFT coefficients. It is constructed so that it can withstand geometric alterations.
• Fourier descriptors: The watermark is embedded in the Fourier descriptors of polygonal lines represented as vector graphics images.
• Self-similar watermarks: A self-similar watermark is embedded in several levels of DWT coefficients.
• Multi-bit watermarks: The watermark that is embedded corresponds to a multiple-bit message and can be decoded even after geometric attacks.

In addition to the above, a theoretic investigation of Markov chaotic sequences and their relative performance in watermarking compared to pseudo-random sequences has been carried out.

Generally, the watermarking scheme is distinguished in three particular algorithms:

a) The Watermark Production algorithm (WPA)
b) The Watermark Embedding algorithm (WEA)
c) The Watermark Detection algorithm (WDA)

The above algorithms are public and can be applied by anyone. The security of the watermark hiding is based exclusively on the privacy of the watermark key.

• WPA is applied having as input the private key and the data of the digital product. Each pair (key, product) is transformed to a watermark digital signal. By passing the data of the product in the algorithm, we manage to achieve the statistical invisibility of the watermark. WPAs are based on pseudorandom number generators or/and strongly chaotic systems. The determination of a key which produces a predefined watermark signal is impossible, i.e. watermark production is a non-invertible procedure preventing an attacker to define counterfeit watermarks. WPA can be considered as a part of the embedding and detection algorithms.
• WEA requires as arguments the digital product and the produced watermark. It embeds the watermark in a digital still image or in a video frame by producing alterations in the luminance of the pixels. For audio signals, alterations are p erformed on the amplitude of the signal. Alterations are of low energy and they are designed by taken into account the main characteristics of the human visual and auditory system. Watermark embedding in transform domains (e.g. the DCT domain) can be used. Since video and audio signals require large amounts of memory, data are input in WEAs and watermarked sequentially. For video watermarking, the motion vectors that calculated by a MPEG encoder-decoder should be used to increase the performance of the algorithm.

Watermarking of digital products

• WDA: Detection is the most crucial part in the watermarking framework. WDAs should be trustworthy producing insignificant false alarm errors. Therefore, the result of the detector should be an indisputable indication about the ownership of a digital product. WDA's based on statistical hypothesis testing. They provide the certainty, which characterises a possible decision that accepts the existence of a particular watermark. Numerical experiments indicate that watermark detection is en ough reliable and suitable to prove copyright ownership. In video frame sequences, the watermark can be detected to each frame separately. WDA does not require the original digital product in order to proceed to the watermark detection. This characteristi c is a great advantage since it provides fast and automatic implementation and "web-crawling''.

Watermark detection

Example of a watermarked image

Downloads

Slide show

Relevant Publications

S. Zafeiriou, A. Tefas and I. Pitas, "Blind Robust Watermarking Schemes for Copyright Protection of 3D Mesh Objects", IEEE Transactions on Visualization and Computer Graphics, vol. 11, no. 5, pp. 596-607, September-October, 2005.

V. Solachidis and I. Pitas, "Watermarking digital 3D volumes in the Discrete Fourier Transform domain", in Proc. of IEEE Int. Conf. on Multimedia and Expo (ICME 2005), Amsterdam, Holland, 6-8 July, 2005.

V. Solachidis, S. Tsekeridou, S. Nikolopoulos and I. Pitas, "Self-Similar watermarks for counterfeiting geometrical attacks", in Proc. of 1st WAVILA CHALLENGE (WaCha 2005), Barcelona, Catalonia (Spain), 8-9 June, 2005.

V. Rodriguez-Doncel, N. Nikolaidis and I. Pitas, "Watermarking Polygonal Lines Using an Optimal Detector on the Fourier Descriptors Domain", in Proc. of 2005 EURASIP European Signal Processing Conf. (EUSIPCO 2005), Antalya, Turkey, 4-8 September, 2005.

V.Solachidis and I.Pitas, "Watermarking polygonal lines using Fourier descriptors", IEEE Computer Graphics and Applications, vol.24, no, 3, pp. 44-51, May-June, 2004.

S. Zafeiriou, A. Tefas and I. Pitas, "A Blind Robust Watermarking Scheme for Copyright Protection of 3D Mesh Models", in Proc. of IEEE Int. Conf. on Image Processing (ICIP 2004), Singapore, 24-27 October, 2004.

A.Nikolaidis and I.Pitas, "Asymptotically Optimal Detection for Additive Watermarking in the DCT and DWT Domains", IEEE Transactions on Image Processing, vol. 12, no. 5, pp. 563-571, May, 2003.

A.Tefas, A.Nikolaidis, N.Nikolaidis, V.Solachidis, S.Tsekeridou and I.Pitas, "Markov chaotic sequences for correlation based watermarking schemes", Chaos, Solitons and Fractals, vol. 17, pp. 567-573, 2003.

A.Tefas, A.Nikolaidis, N.Nikolaidis, V.Solachidis, S.Tsekeridou, and I.Pitas, "Performance Analysis of Correlation-Based Watermarking Schemes Employing Markov Chaotic Sequences", IEEE Transactions on Signal Processing, vol. 51, no. 7, pp. 1979-1994, July, 2003.

A. Gianoula, A. Tefas, N. Nikolaidis and I. Pitas, "Improving the Detection Reliability of Correlation-based Watermarking Techniques", in Proc. of IEEE Int. Conf. on Multimedia and Expo 2003(ICME2003), Baltimore, Maryland, USA, 6-9 July, 2003.

A. Kalivas, A. Tefas and I.Pitas, "Watermarking of 3D models using Principal Component Analysis", in Proc. of 2003 IEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP 2003) and in Proc. of 2003 IEEE Int. Conf. on Multimedia and Expo (ICME2003), Baltimore, Maryland, USA, 6-9 July, 2003.

S. Zafeiriou, A. Kalyvas, A. Tefas, N. Nikolaidis and I. Pitas, "Watermarking of 3D mesh models robust to basic geometric transformations", in Proc. of 9th Panhellenic Conf. on Informatics (PCI `03), Thessaloniki, Greece, 21-23 November, 2003.

A. Giannoula, N. Nikolaidis and I. Pitas, "Watermarking of Sets of Polygonal Lines using Fusion Techniques", in Proc. of IEEE Int. Conf. on Multimedia and Expo 2002 (ICME2002), vol. 2, pp. 549-552, Laussane, Switzerland, 26-29 August, 2002.

N. Nikolaidis, V. Solachidis, A. Tefas, V. Arguriou and I. Pitas, "Benchmarking of Still Image Watermarking Methods: Principles and State of the Art", in Proc. of Electronic Imaging & the Visual Arts 2002 (EVA2002), Florence, Italy, 18-22 March, 2002.

N. Nikolaidis, V. Solachidis, A. Tefas and I. Pitas, "Watermark Detection: Benchmarking Perspectives", in Proc. of IEEE Int. Conf. on Multimedia and Expo 2002(ICME2002), Special Session on Benchmarking of Data Hiding Technologies, vol. 2, pp. 493-496, Lausanne, Switzerland, 26-29 August, 2002.

A. Nikolaidis and I.Pitas, "Optimal Detector Structure for DCT and Subband Domain Watermarking", in Proc. of IEEE Int. Conf. on Image Processing 2002 (ICIP2002), vol. 3, pp. III-465 - III-468, Rochester, N.Y., USA, 22-25 September, 2002.

A. Tefas, G. Louizis and I. Pitas, "3D Image Watermarking robust to Geometric Distortions", in Proc. of IEEE Int. Conf. on Acoustics Speech and Signal Processing 2002 (ICASSP2002), vol. 4, pp. 3465-3468, Orlando, Florida, USA, 13-17 May, 2002.

S. Tsekeridou, V. Solachidis, N. Nikolaidis, A. Nikolaidis, A. Tefas and I. Pitas, "Statistical Analysis of a Watermarking System based on Bernoulli Chaotic Sequences", Signal Processing, Elsevier Special Issue on Information Theoretic Issues in Digital Watermarking, 81 (6), pp. 1273-1293, 2001.

F. Bartollini, A. Tefas, M. Barni and I. Pitas, "Image Authentication Techniques for Surveillance Applications", Proceedings of the IEEE , vol. 89, no. 10, pp. 1403-1418, September, 2001.

P. Bassia, I. Pitas, N. Nikolaidis, "Robust audio watermarking in the time domain", IEEE Transactions on Multimedia, vol 3 no. 2, pp. 232 -241, June 2001.

V. Solachidis and I. Pitas, "Circularly Symmetric Watermark Embedding in 2D DFT domain", Transactions on Image Processing, vol. 10, no. 11, pp. 1741-1753, November, 2001.

A.Nikolaidis and I.Pitas, "Region-Based Image Watermarking", IEEE Transactions on Image Processing, vol. 10, no. 11, pp. 1726-1740, November, 2001.

A. Nikolaidis, S. Tsekeridou, A. Tefas and V. Solachidis, "A survey on watermarking application scenarios and related attacks", IEEE International Conference on Image Processing (ICIP 2001), Thessaloniki, Greece, 7-10 October 2001.

V. Solachidis, A. Tefas, N. Nikolaidis, S. Tsekeridou, A. Nikolaidis, I.Pitas, "A benchmarking protocol for watermarking methods", 2001 IEEE Int. Conf. on Image Processing (ICIP'01), Thessaloniki, Greece, 7-10 October, 2001.

V. Solachidis, S. Tsekeridou, A. Tefas, I.Pitas, "A simple benchmarking system for performance evaluation of still image watermarking methods", 1st International Symposium on Telecommunications (IST'01), accepted for publication, Tehran, IRAN; 1-3 September, 2001.

A. Tefas , A. Nikolaidis, N. Nikolaidis, V. Solachidis, S. Tsekeridou and I. Pitas, "Statistical Analysis of Markov Chaotic Sequences for Watermarking applications", 2001 IEEE International Symposium on Circuits and Systems (ISCAS 2001), CD-ROM, May 6 - 9, Sydney, Australia, 2001.

A. Tefas and I. Pitas, "Robust Spatial Image Watermarking using Progressive Detection", Proc. of 2001 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2001), CD-ROM, Salt Lake City, Utah, USA, 7-11 May 2001.

A. Tefas , A. Nikolaidis, N. Nikolaidis, V. Solachidis, S. Tsekeridou and I. Pitas, "Performance Analysis of Watermarking Schemes based on Skew Tent Chaotic Sequences", NSIP'01, CD-ROM, Baltimore, Maryland, USA, 3-6 June, 2001.

A. Tefas , A. Nikolaidis, N. Nikolaidis, V. Solachidis, S. Tsekeridou and I. Pitas, "Markov Chaotic Sequences for Correlation based Watermarking schemes", International Conference on Nonlinear Dynamics (NLD'01), Thessaloniki, Greece, 27-30 August, 2001.

S. Tsekeridou, V. Solachidis, N. Nikolaidis, A. Nikolaidis, A. Tefas, I. Pitas, "Bernoulli Shift Generated Chaotic Watermarks: Theoretic Investigation", IEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP2001), Salt Lake City, Utah, USA, 7-11 May 2001.

S. Tsekeridou, V. Solachidis, N. Nikolaidis, A. Nikolaidis, A. Tefas, I. Pitas, "Theoretic Performance Analysis of a Watermarking System based on Bernoulli Chaotic Sequences", Communications and Multimedia Security Conf. (CMS 2001), Darmstadt, Germany, 21-22 May 2001.

S. Tsekeridou, V. Solachidis, N. Nikolaidis, A. Nikolaidis, A. Tefas, I. Pitas, "Theoretic Investigation of the Use of Watermark Signals derived from Bernoulli Chaotic Sequences", 12th Scandinavian Conference on Image Analysis 2001 (SCIA2001), accepted for publication, Bergen, Norway, 11-14 June 2001.

A. Nikolaidis and I. Pitas, "Comparison of different chaotic maps with application to image watermarking", IEEE Int. Symposium on Circuits and Systems (ISCAS 2000), Geneva, Switzerland, vol. V, pp. 509-512, 28-31 May 2000.

A. Nikolaidis and I. Pitas, "A region-based technique for chaotic image watermarking", EURASIP European Signal Processing Conference (EUSIPCO 2000), Tampere, Finland, 5-8 September 2000.

N. Nikolaidis, S. Tsekeridou, A. Nikolaidis, A. Tefas, V. Solachidis and I. Pitas, "Applications of chaotic signal processing techniques to multimedia watermarking", Proceedings of the IEEE workshop on Nonlinear Dynamics in Electronic Systems, pp. 1-7, Catania Italy, May 18-20 2000.

V. Solachidis, N. Nikolaidis and I. Pitas , "Watermarking Polygonal Lines Using Fourier Descriptors", IEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP'2000), Istanbul, Turkey, vol. IV, pp 1955-1958, 5-9 June 2000.

V. Solachidis, I. Pitas, "Self-similar ring shaped watermark embedding in 2-D DFT domain", 2000 European Signal Processing Conf. (EUSIPCO'00) CD-ROM, Tampere, Finland, 5-8 September 2000.

V. Solachidis, N. Nikolaidis, I. Pitas, "Fourier descriptors watermarking of vector graphic images", 2000 IEEE Int. Conf. on Image Processing (ICIP'00) accepted for publication, Vancouver, Canada, 10-13 September 2000.

A. Tefas and I. Pitas, "Multi-bit Image Watermarking Robust to Geometric Distortions", IEEE International conference on Image Processing (ICIP'2000) CD-ROM, Vancouver, Canada, 10-13 September 2000.

A. Tefas and I. Pitas, "Image Authentication based on Chaotic Mixing", IEEE International Symposium on Circuits and Systems (ISCAS'2000), Geneva, Switzerland, 28-31 May 2000.

A. Tefas and I. Pitas, "Image Authentication and Tamper Proofing using Mathematical Morphology", European Signal Processing Conference (EUSIPCO'2000), Tampere, Finland, 5-8 September, vol.3, pp. 1681-1684, 2000.

S. Tsekeridou, I. Pitas, "Wavelet-based Self-Similar Watermarking for Still Images", 2000 IEEE Int. Symposium on Circuits and Systems (ISCAS'00) vol. I, pp. 220-223, Geneva, Switzerland, 28-31 May 2000.

S. Tsekeridou, I. Pitas, "Embedding Self-Similar Watermarks in the Wavelet Domain", 2000 IEEE Int. Conf. on Acoustics, Systems and Signal Processing (ICASSP'00), vol. IV, pp. 1967-1970, Istanbul, Turkey, 5-9 June 2000.

S. Tsekeridou, N. Nikolaidis, N. Sidiropoulos, I. Pitas, "Copyright Protection of Still Images Using Self-Similar Chaotic Watermarks", 2000 IEEE Int. Conf. on Image Processing (ICIP'00), Vancouver, Canada, 10-13 September 2000.

A. Nikolaidis and I. Pitas, "Robust watermarking of facial images based on salient geometric pattern matching", IEEE Trans. on Multimedia vol. 2, no. 3, pp. 172-184, September 2000.

S. Stankovic, I. Djurovic and I. Pitas, "Watermarking in the Space/Spatial Domain Using Two-Dimentional Radon Wigner distribution", IEEE Trans. on Image Processing, 2000.

A. Nikolaidis and I. Pitas, "A Robust Feature-based Technique for Watermarking Frontal Face Images", IEEE-EURASIP Wor. on Nonlinear Signal and Image Processing (NSIP'99), vol. 1, pp. 341-345, Antalya, Turkey, 20-23 June 1999.

N. Nikolaidis and I. Pitas, "Digital Image Watermarking: an Overview", Int. Conf. on Multimedia Computing and Systems (ICMCS'99), vol. I, pp. 1-6, Florence, Italy, 7-11 June 1999.

V. Solachidis and I. Pitas, "Circularly symmetric watermark embedding in 2-D DFT domain", IEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP'99), Phoenix, Arizona, USA, vol.6, pp. 3469-3472, 15-19 March 1999.

C. Busch, K. Nahrstedt and I. Pitas, "Image Security", IEEE Computer Graphics and Applications, vol. 19, no. 1, pp. 16-17, Jan.-Feb. 1999.

G. Voyatzis and I. Pitas, "Problems and Challenges in Multimedia Networking and Content Protection", TICSP Series #3, Trends and Challenges, Contributions to Workshop on Trends and Important Challenges in Signal Processing, ed. Jaakko Astola, March 1999.

G. Voyatzis and I. Pitas, "Protecting digital image copyrights: a framework", IEEE Computer Graphics and Applications, vol. 19, no. 1, pp. 18 -24, Jan.-Feb. 1999.

G. Voyatzis and I. Pitas, "The use of watermarks in the protection of digital multimedia products", Proceedings of the IEEE, vol. 87, no. 7, pp. 1197-1207, July 1999.

G. Voyatzis, N. Nikolaidis and I. Pitas, "Copyright Protection of Multimedia Documents: From Theory to Application", 4th Hellenic European Conference on Computer Mathematics and its Applications, 1998.

A.G. Bors and I. Pitas, "Image Watermarking Using Block Site Selection and DCT Domain Constraints", Optics Express, vol. 3, no. 12, pp. 512-522, Dec. 1998.

N. Nikolaidis and I. Pitas, "Robust image watermarking in the spatial domain", Signal Processing, Elsevier, vol. 66, no. 3, pp. 385-403, 1998.

I. Pitas, "A Method for Watermark Casting in Digital Images", IEEE Trans. on Circuits and Systems on Video Technology, vol. 8, no.6, October 1998.

G. Voyatzis and I. Pitas , "Digital Image Watermarking using Mixing Systems", in Computer & Graphics, vol. 22, no. 3, 1998.

P. Bassia and I. Pitas, "Robust Audio Watermarking in the Time Domain", in Proc. of IX European Signal Processing Conf. (EUSIPCO'98), vol. 1, pp. 25-28, Rhodes, Greece, 8-11 September, 1998.

G. Voyatzis, N. Nikolaidis and I. Pitas, "Digital Watermarking: An Overview", in Proc. of IX European Signal Processing Conf. (EUSIPCO'98), vol. 1, pp. 9-12, Rhodes, Greece, 8-11 September, 1998.

G. Voyatzis and I. Pitas, "Chaotic Watermarks for Embedding in the Spatial Digital Image Domain", in Proc. of IEEE Int. Conf. on Image Processing (ICIP'98), vol. 2, pp. 432-436, Chicago, Illinois, USA, 4-7 October, 1998.

G. Voyatzis and I. Pitas, "Embedding Robust Watermarks by Chaotic Mixing", 13th International Conference on Digital Signal Processing (DSP'97), Santorini, Greece, vol. 1, pp. 213-216, 2-4 July 1997.

A. Bors and I. Pitas, "Embedding Parametric Digital Signatures in Images", EUSIPCO-96, Trieste, Italy, vol. III, pp. 1701-1704, September 1996.

A. Bors and I. Pitas, "Image Watermarking Using DCT Domain Constraints", 1996 IEEE International Conference on Image Processing (ICIP'96), Lausanne , Switzerland, vol. III, pp. 231-234, 16-19 September 1996.

N. Nikolaidis and I. Pitas, "Copyright protection of images using robust digital signatures", IEEE International Conference on Acoustics, Speech a nd Signal Processing (ICASSP-96), vol. 4, pp. 2168-2171, May 1996.

I. Pitas, "A Method for Signature Casting on Digital Images", 1996 IEEE International Conference on Image Processing (ICIP'96), Lausanne, Switzerland, vol. III, pp. 215-218, 16-19 September 1996.

G. Voyatzis and I. Pitas, "Chaotic Mixing of Digital Images and Applications to Watermarking", European Conference on Multimedia Applications, services and Techniques (ECMAST'96), Louvain-la-Neuve, Belgium, vol. 2, pp. 687-695, May 1996.

G. Voyatzis and I. Pitas, "Applications of Toral Automorphisms in Image Watermarking", 1996 IEEE International Conference on Image Processing (ICIP'96), Lausanne, Switzerland, vol. II, pp. 237-240, 16-19 September 1996.

I. Pitas and T. H. Kaskalis, "Applying Signatures on Digital Images", IEEE Workshop on Nonlinear Image and Signal Processing, Neos Marmaras, Greece, pp. 460-463, June 1995.

Research Projects

ECRYPT - European Network of Excellence in Cryptology, IST, FP6

CERTIMARK - Certification for watermarking techniques, IST, EC

INSPECT - Innovative Signal Processing Exploiting Chaotic Dynamics, LTR-ESPRIT, EC

OKAPI - Open Kernel for Access to Protected Interoperable interactive services, ACTS, EC

ACCOPI - Access control and copyright protection for images, RACE, EC

Archimedes - Research Group Support in TEI (EEOT)

One of the most fundamental technologies necessary for the management of digital video is the retrieval (from a video database) of one or more video segments that the user is interested in. The methods used for approaching video retrieval are similar to those used for the retrieval of other types of multimedia objects, such as images. Retrieval usually follows one of two paradigms:

• Query-by-keyword: A video database is annotated with keywords or other metadata. The user then enters the keywords that best describe what he is searching for or some other appropriate metadata. These metadata are then used to perform a textual or symbolic search in the database.
• Query-by-example: The videos in a database are characterized with an appropriate set of features, which constitute a representation of the digital item. We call this representation a signature . The user then inputs a video similar to the one that he is searching for. Then, a set of features is extracted from the user video and used to find images or videos with similar features.

Another technology which is useful for the management of video, particularly with respect to rights protection, is fingerprinting. This is defined as the identification of a video segment using a representation called fingerprint , which is extracted from the video content. The fingerprint must uniquely identify a video segment, and must be invariant to manipulation of the video.

Our Method

The general idea of our approach is that the existence of faces of specific individuals can be used to characterize a video segment. Assuming that the faces in the video have been detected and identified, the video signature (or fingerprint) consists of quartets of the following:

• Appearance time/frame a
• Disappearance time/frame b
• Identity of person s(denoted below by color)
• Certainty of appearance F

Given the above representation, we compute the similarity of two videos, for a certain displacement d:

where Fi(n,m) is the certainty that person m appears in frame n of video segment i.

Using this representation, our algorithm for retrieval is as follows:

1. Find, in the query segment, the quartet with the greatest area.
2. Find, through an index, all database quartets that refer to the same person as the above quartet.
3. Calculate, for each quartet found, the range of displacements that can result in a match



4. Match query quartets with compatible (same person) quartets in the database in that range of displacements
5. For each such pair, compute the area of the overlap of the pulses
6. The optimal matching location is at the maximum overall overlap

Downloads

-

Relevant Publications

C. Cotsaces, N. Nikolaidis and I. Pitas, "The use of face indicator functions for video indexing and fingerprinting", in Proc. of Int. Workshop on Content-Based Multimedia Indexing (CBMI 2005), Riga, Latvia, 21-23 June, 2005.

C. Cotsaces, N.Nikolaidis and I.Pitas, "Video Indexing by Face Occurrence-based Signatures", in Proc. of 2006 IEEE Int. Conf. on Acoustics, Speech and Signal Processing, Toulouse, France, 14-19 May, 2006.

Research Projects

MUSCLE - “Multimedia Understanding through Semantics, Computation and LEarning” (FP6-507752)

In recent years, digital technology, computers and networks have become a must in our everyday life. The power of personal computers is getting stronger and stronger as time moves on. Huge numbers of multimedia data transmissions are passing through the internet. Consequently, to pirate these data is easy and takes no time. Therefore, how to protect our digital data from being copied and used illegally has become a major concern.

A well-known solution to protecting the copyright of images is digital watermarking. Digital watermarking is a technique which allows an individual to add hidden copyright notices or other verification messages to digital audio, video, or image signals and documents. Such hidden message is a group of bits describing information pertaining to the signal or to the author of the signal (name, place, etc.). The technique takes its name from watermarking of paper or money as a security measure. Digital watermarking can be a form of steganography, in which data is hidden in the message without the end user's knowledge.

Another well known solution is the use of fingerprints. Fingerprints are perceptual features or short summaries of a multimedia object. This features stored in a vector called feature vector. This feature vector characterizes unique the multimedia object. Feature vectors can be stored to a database and later on can be used to trace a copy of the original multimedia object.

Our Method

At present the research is focused on digital images. A color based approach is used to extract some unique characteristics of each image that later on can be used to identify it uniquely. Based on research and experimental results a limited chromatic histogram (24 colours) used to describe the image and a set of 24 values for each image stored into a data structure (R-Tree). R-trees are used because are fast in multidimensional searches.

The fingerprint extraction procedure involves the quantization of the image colors and the calculation of color histograms based on the resulting colors. We used a quantization method based on a pre-defined color palette known as the Gretag Macbeth Color Checker, which was designed to match human perception of colors. The Macbeth chart is a standard used to test color reproduction systems and it consists of 24 colors, scientifically prepared to represent a variety of different naturally occurring colors.

A database of 450 art images of variable sizes, provided by the Bridgeman Art Library, was used to evaluate the method. The following set of 20 transformations was applied to each image: Scaling (25, 50, 75, 125, 150 and 200%), Rotation (10, 20, 30, 90), Cropping - (both sides by 10, 20 and 30%), Compression - (JPEG with 25, 50, 75 quality factor), Blurring (median with 3x3, 5x5,7x7 masks), and Combination of attacks (Rotation 10, cropping 10%, resizing to 25%, median filtering 5x5 and compression with quality factor 50). The images were resized using nearest neighbor interpolation.

For the image rotation, it has to be noted that a black frame was added around the image, thus producing an additional source of degradation, except for the case of 90. The set of 450 original images defined the Original Image Set and the resulting set of 9000 transformation images defined the Transformed Image Set. The use of color-based descriptors for the application of image fingerprinting was evaluated using Receiver-OperatorCharacteristic (ROC) analysis. Specifically, the evaluation consisted of taking the color descriptors from each of the 450 images in the Original Image Set and matching them against the descriptors from each of the 9000 images in the Transformed Image Set. Matches were determined by applying a threshold on the similarity measures and identifying those images with measures higher than the threshold. The well-known measures True Positive Fraction (TPF or sensitivity) and False Positive Fraction (FPF) were used. By sweeping the threshold and averaging the measures of TPF and FPF over all images in the Original Image Set, ROC curves were calculated.

It can be seen, that the normalized color histogram with the similarity measures scaled L 1-norm and scaled histogram intersection show the best performance whereas the quadratic histogram measure shows the worst performance. The quadratic histogram measure incorporates information regarding the distance of colors in the color space, which might be more useful if the query was for images of similar color, as opposed to exact matches. It can also be seen that the spatial chromatic histogram shows slightly worse performance compared to the color-only histogram for this experiment. The spatial information could prove useful when two images have exactly the same colors but in different locations. However,we did not design a database having those requirements since the goal was to examine the robustness of color-based descriptors over transformation changes in a general database.

Based on the above results a large image database system created that extracts and stores fingerprints and images. Using this system, we can identify copies of an original, copyright protected image even if in that copy has been made various attacks. In more detail, the system can accept an image as input and return as a result the exactly same image - if found- or a set of possible same images that are close enough to the original question image. The system has many functions, including a full -user friendly- web interface, functions for insert, modify and delete images, indexing of the stored images so the search later on can be done fast, various methods for comparing the input image with each one of the database images and many more.

A screenshot from the main search page of the system

Search results

Future research contains more study on the above techniques and even faster search using advanced multidimensional indexing methods, so the method can be used in other multimedia data (e.g. video) as well as in image data.

Downloads

-

Relevant Publications

M. A. Gavrielides, E. Sikudova, D. Spachos, I. Pitas, "Color Features for Image Fingerprinting", SETN 2006, Heraklion, Greece, 18-20 May, 2006.

Research Projects

Archimedes - Research Group Support in TEI (EEOT)