In this article we present a semi-fragile watermarking scheme for authenticating intra-coded frames in compressed digital videos. The scheme provides the detection of content-changing manipulations while being moderately robust against content-preserving manipulations. More generally, we mean by content-preserving manipulations those, which are applied in post-production processes, such as compression. Content-changing manipulations remove or insert objects into frames or sequences of frames. We focus in this work on a semi-fragile watermarking method based on invariant features referred to as points of interests. The features are extracted using the Moravec-Operator. The interest point operator of Moravec is totally un-supervised and does not require any a priori knowledge in the class of objects being protected in a given frame. Out of the interest points we generate a binary mask, which will be embedded robustly as watermark into the video. In the verification process we compare the detected watermark with the points of interest from the video, which has to be verified. We present test results evaluating the robustness against content-preserving manipulations and the fragility regarding content-changing manipulations. Beside the discussion of the results we propose a procedure to provide security of the scheme against forgery attacks.
In this paper, we present our work for automatic generation of textual metadata based on visual content analysis of video news. We present two methods for semantic object detection and recognition from a cross modal image-text thesaurus. These thesaurus represent a supervised association between models and semantic labels. This paper is concerned with two semantic objects: faces and Tv logos. In the first part, we present our work for efficient face detection and recogniton with automatic name generation. This method allows us also to suggest the textual annotation of shots close-up estimation. On the other hand, we were interested to automatically detect and recognize different Tv logos present on incoming different news from different Tv Channels. This work was done jointly with the French Tv
Channel TF1 within the "MediaWorks" project that consists on an hybrid text-image indexing and retrieval plateform for video news.