Program

Thursday 17th of February

10.30-11.00 Coffee and croissants, registration desk open

11.00-12.30 Local features, large scale search and classification by Dr. Cordelia Schmid

12.30-14.00 Lunch

14.00-15.30 Local features, large scale search and classification by Dr. Cordelia Schmid

15.30-15.45 Coffee break

15.45-17.00 Open panel

Friday 18th of February

10.30-11.00 Coffee and croissants, registration desk open

11.00-12.30 Privacy-preserving signal processing by prof. Inald Lagendijk

12.30-14.00 Lunch

14.00-15.30 Privacy-preserving signal processing by prof. Inald Lagendijk

15.30-15.45 Coffee break

15.45-17.00 Open panel

 

Abstracts

Local features and large scale image search

Cordelia Schmid, LEAR team, INRIA Grenoble

In this presentation we address the problem of image search on a large scale. This requires a robust image representation as well as a rapid access mechanism.

We, first, introduce a robust image representation based on local features. Local features are distinctive and can operate in the presence of large scale and viewpoint changes as well as occlusion and clutter. We present detectors for scale and affine invariant local regions, such as the Hessian [1] and MSER [2] detectors, as well as their description, for example with the SIFT descriptor [3]. We, then, show how to use these features for robust image matching. To find corresponding images in a large collection, we introduce the bag-of-features framework [4] which permits a rapid access by quantizing the features and using an inverted file system. We, also, present recent improvements of the bag-of-features approach, such as Hamming embedding [5]. We conclude our talk by presenting methods for very large scale search, where each image is represented by a single descriptor, such as the Fisher vector [6], which is, then, compressed into a very compact code [7].

[1] K. Mikolajczyk and C. Schmid. Scale and affine invariant interest
point detectors. International Journal of Computer Vision, 2004.

[2] J. Matas, O. Chum, M. Urba, and T. Pajdla. Robust wide baseline
stereo from maximally stable extremal regions. British Machine Vision
Conference, 2002.  

[3] D. Lowe. Distinctive image features from scale-invariant
keypoints. International Journal of Computer Vision, 2004. 

[4] J. Sivic and  A. Zisserman. Video Google: A text retrieval
approach to object matching in videos. International Conference on
Computer Vision, 2003. 

[5] H. Jegou, M. Douze and C. Schmid. Hamming embedding and weak
geometric consistency for large-scale image search. European
Conference on Computer Vision, 2008.   

[6] F. Perronnin, Y. Liu, J. Sanchez and H. Poirier. Large-Scale image
retrieval with compressed Fisher vectors. Conference on Computer
Vision and Pattern Recognition, 2010.

[7] H. Jegou, M. Douze, C. Schmid and P. Perez. Aggregating local
descriptors into a compact image representation. Conference on
Computer Vision and Pattern Recognition, 2010.

Privacy-Preserving Signal Processing

Prof. R. (Inald) Lagendijk, Delft University of Technology.

In this presentation I will discuss the problems, principles and examples of protecting the privacy of users in multimedia applications. Some multimedia applications pose serious privacy threats for their users as they rely on privacy-sensitive information that can be misused. To protect the privacy of users, an emerging paradigm shows that it is attractive and feasible to combine signal processing and cryptography [1].

The focus is on those applications that are executed remotely or “in the cloud”, such as on-line recommendation services (amazon.com, Google.com) but also face-recognition systems. The new and exciting research area of privacy-preserving signal processing aims at making privacy-sensitive data of the user of such multimedia applications inaccessible by means of encryption. Although it is then impossible for the service provider to access directly the content of the encrypted data without the decryption key, the service provider can still process the data under encryption to perform the required task. The protocols to process the encrypted data are designed by using cryptographic primitives like homomorphic cryptosystems [2] and secure multiparty computation techniques [3].

The first part of the presentation starts with an overview of a number of relevant privacy-preserving signal processing cases and related security requirements. Typical signal processing operations such as linear transforms will be discussed briefly. Next the prototypical cryptographic protocols will be introduced in the context of the signal processing operations. These protocols include blinding, multi-party computation, and homomorphic crypto systems. In the second part, I describe a number of examples that show solutions for privacy-preserving signal processing, including privacy-preserving face recognition and secure clustering [4,5,6]. Several open issues in “designing privacy in signal processing” are finally presented for further discussion.

[1] Protection and retrieval of encrypted multimedia content: when cryptography meets signal processing; Zekeriya Erkin, Alessandro Piva, Stefan Katzenbeisser, Reginald L. Lagendijk, Jamshid Shokrollahi, Gregory Neven, and Mauro Barni; EURASIP Journal on Information Security, Volume 2007, p.20 (2007).

[2] A survey of homomorphic encryption for nonspecialists; Caroline Fontaine and Fabien Galand; EURASIP Journal on Information Security, Volume 2007 (2007).

[3] Foundations of Cryptography: Volume 2, Basic Applications; Oded Goldreich; Cambridge University Press, New York, NY, USA (2004).

[4] Privacy-Preserving User Clustering in a Social Network; Zekeriya Erkin, Thijs Veugen, Tomas Toft and Reginald L. Lagendijk; IEEE International Workshop on Information Forensics and Security (2009).

[5] Privacy Enhanced Recommender System; Zekeriya Michael Beye, Thijs Veugen and Reginald L. Lagendijk; Thirty-first Symposium on Information Theory in the Benelux, Rotterdam (2010).

[6] Privacy-Preserving Face Recognition, Zekeriya Erkin, Martin Franz, Jorge Guajardo, Stefan Katzenbeisser, Reginald L. Lagendijk, and Tomas Toft, 9th International Symposium on Privacy Enhancing Technologies, August, p.235-253 (2009).