Catálogo de publicaciones - libros

With the rising popularity of the real-time audio streaming applications, it is important, in order to support (and to charge) properly these new services, to understand and characterize the behaviour of these new applications. This paper studies the traffic associated with a specific audio streaming on-live service, the radio over the Internet, and presents some analyses and comparisons of different streaming media products. In addition, the user’s viewpoint is also taken into account, since it is ultimately the users (and their degree of satisfaction) who will decide whether a service goes to be or not successful. Finally, the study estimates the real bandwidth used by the on-line radio and examines how a cable network based on DOCSIS, have to be planned in order to be able to support this new service.

Energy consumption is a key issue in modern microprocessor system design in general, and in the design of mobile computing devices more in particular. This paper introduces a novel approach to energy-efficient media stream decoding that is based on the notion of media stream similarity. The key idea is that platform-independent scenarios of similar decode complexity can be identified within and across media streams. A client decoding a media stream annotated with scenario information can then adjust its processor clock frequency and voltage level based on these scenarios for reduced energy consumption. Our evaluation done using the AVC decoder and 12 reference streams shows an average energy reduction of 46% while missing less than 0.2% of the frame deadlines on average.

We investigate optimization of system performance in the below scenario: capturing and transmitting videos by single or multiple video sensors using distributed video coding (DVC) over ad hoc networks. There is an intrinsic contradiction in this scenario that could affect the system performance: the contradiction between the decoding quality and network lifetime. In this paper, we propose a joint optimization between the decoding quality and network lifetime using a quantitative metric of system performance, which is defined as the amount of collected visual information during the operational time of the video sensor. Based on the proposed metric, an optimal encoding rate is determined, which results in an optimal system performance. The simulation results show that the optimal encoding rate can be determined to achieve the optimal system performance.

In conventional video coding systems, the encoder performs predictive coding (motion estimation) to exploit the temporal similarity, which make its complexity much higher than that of the decoder. The Wyner-Ziv theory or Lossy Distributed Coding theory suggests that when frames are separately encoded but jointly decoded, similar coding efficiency could be achieved. Trellis coded quantization (TCQ) is a powerful quantization method which has already shown its power in distributed source coding area. Recently, some applicable Wyner-Ziv video coding systems have been proposed. In this paper, based on the PRISM system, TCQ is employed to improve the performance. In order to minimize the complexity increase, only 4-state trellis is used, and no codebook is trained or stored, and we also propose a method to solve the refinement problem. Some other changes are also made to further improve the performance. Experimental results indicate that though the simplest TCQ is used, gain is also achieved over scalar quantization.

In this paper, an efficient VLSI architecture of full-search variable block size motion estimation (VBSME) suitable for high quality video is proposed. Memory bandwidth in high-quality video is a mainly responsible for throughput limitations and power consumption in VBSME. The proposed architecture is designed for reducing the memory bandwidth by adopting “meander”-like scan for a high overlapped data of the search area and using on-chip memory to reuse the overlapped data. We can reuse the previous candidate block of 98% for the current one and save memory access cycles about 19% in a search range of [-32, +31]. The architecture has been prototyped in Verilog HDL and synthesized by Synopsys Design Compiler with Samsung 0.18um standard cell library. Under a clock frequency of 67MHz, The simulation result shows that the architecture can achieve the real-time processing of 720x576 picture size at 30fps with the search range of [-32~+31].

The purpose of an image hash is to provide a compact representation of the whole image. Designing a good image hash function requires careful consideration of many issues such as robustness, security and tamper detection with precise localization. In this paper, we present a novel hashing scheme that addresses these issues in a unified framework. We analyze the security issues in image hashing and present new ideas to counter some of the attacks that we shall describe in this paper. Our proposed scheme is resilient to allow non-malicious manipulations like JPEG compression, high pass filtering and is sensitive enough to detect tampering with precise localization. Several experimental results are presented to demonstrate the effectiveness of the proposed scheme.

TV logo detection, tracking and removal play an important role in the applications of claiming video content ownership, logo-based broadcasting surveillance, commercial skipping, and program rebroadcasting with new logos. In this paper, we present a novel and robust framework using tensor method for these three tasks. First, we use tensor based generalized gradient and the OTSU binarization algorithm to logo detection, and propose a two level framework from coarse to fine to tracking the TV logos. Finally, we extend the regularization PDEs by incorporation of temporal information to inpaint the logo region. Due to the introduction of the structure tensor, the generalized gradient based method can detect the logo region by tracking the change rate of pixels in spatio-temporal domain, and the region of logo removal is well filled in a structure-preserving way. Since temporal correlation of multiple consecutive frames is considered, the proposed method can deal with opaque, semi-transparent, and animated logos. The experiments and comparison with previous methods are conducted on the part of TRECVID 2005 news corpus and several Chinese TV channels with challenging TV logos, and the experimental results are promising.

This paper presents a system for automatic detecting and recognizing complex individual actions in sports video to facilitate high-level content-based video indexing and retrieval. This is challenging due to the cluttered and dynamic background in sports video which makes object segmentation formidable. Another difficulty is to fully automatically and accurately detect desired actions from long video sequence. We propose three techniques to handle these challenges. Firstly, an efficient approach exploiting dominant motion and semantic color analysis is developed to detecting the highlight clips which contain athlete’s action from video sequences. Secondly, a robust object segmentation algorithm based on adaptive dynamic background construction is proposed to segment the athlete’s body from the clip. Finally, to recognize the segmented body shape sequences, the hidden markov models are slightly modified to make them suitable for noisy data processing. The proposed system for broadcast diving video analysis has achieved 96.6% detection precision; and 85% recognition accuracy for 13 kinds of diving actions.

In this paper, we propose two novel semantic event detection models, i.e., Two-dependence Bayesian Network (2d-BN) and Conditional Random Fields (CRFs). 2d-BN is a simplified Bayesian Network classifier which can characterize the feature relationships well and be trained more efficiently than traditional complex Bayesian Networks. CRFs are undirected probabilistic graphical models which offer several particular advantages including the abilities to relax strong independence assumptions in the state transition and avoid a fundamental limitation of directed probability graphical models. Based on multi-modality fusion and mid-level keywords representation, we use a three-level framework to detect semantic events. The first level extracts audiovisual features, the mid-level detects semantic keywords, and the high-level infers events using 2d-BN and CRFs models. Compared with state of the art, extensive experimental results demonstrate the effectiveness of the proposed two models.

The explosive growth of information technology and digital content industry stimulates various video applications over the Internet. Since it is quite easy to copy, reformat, modify and republish video files on the websites, similarity/duplicate detection and measurement is essential to identify the excessive content duplication, so as to facilitate effective video search and intelligence propriety protection as well. In this paper, we propose a novel signature-based approach for duplicate video comparison. The so-called scheme counts the numbers of video’s frames that are closest to a set of representative seed vectors chosen from the feature space of the training data set in advance. Then all the numbers are normalized to generate the signature of the video for further comparison. As our signature is a compact fixed-size vector with low dimension for each video, it requires less storage and computation cost than previous methods. The experiments show that our approach is both efficient and effective for web video duplicate detection.