The on-going improvement of the digital image processing technologies together with the usage of specialised hardware has lead to the development of the system we are now introducing. The system incorporates a videose...
详细信息
The on-going improvement of the digital image processing technologies together with the usage of specialised hardware has lead to the development of the system we are now introducing. The system incorporates a videosensor with improved capabilities on nuisance alarm discrimination based on an algorithm which combines both movement detection and object tracking with the aim of providing a wide range of criteria to program and select alarm situations. A TMS320C44 digital signal processor from Texas Instruments Inc. carries out the videosensing algorithm, being able to process up to four different and simultaneous video sources (CCTV cameras). The algorithm allows the usage of both local and global parameters to be applied to particular image areas or to the whole video frame. This set of parameters include: maximum speed of the moving objects; localisation of the areas of interest (alarm zones); order of intersection of the moving objects with the alarm zones or the size of the objects along their trajectories. The other basic component of the system is based on a video codec (IIT VCP) which uses the H.263 ITU-T standard on real time video coding. The high compression rates which the H.263 accomplishes allows the transmission of real time video through low bandwidth data links (28800 to 64000 bps) and the utilisation of the IP protocol makes it feasible to integrate our system in practically any data network with a minimal investment in communication equipment.
video coding is used for lot of multimedia purposes like video conferencing, digital storage media, Internet streaming and television broadcasting. This paper presents a new design for the implementation of Full-Searc...
详细信息
video coding is used for lot of multimedia purposes like video conferencing, digital storage media, Internet streaming and television broadcasting. This paper presents a new design for the implementation of Full-Search (FS) Variable Block Size (VBS) Motion Estimation (ME), which is a key issue of different videocompression standards such as MPEG-1, MPEG-2, MPEG-4 Visual, H.261, H.263 and H.264. The FS algorithm is widely used for implementation of ME in videocompressionalgorithms. This design is fully parametric in terms of block size, which is variable, and the Sum of Absolute Differences (SAD) is presented by re-using the outputs. The design features high efficiency in terms of operating frequency and reduction in hardware complexity. These architectures are designed using Verilog Hardware Description Language (HDL) and the functionalities are verified using ModelSim Simulator. For two different designs, namely 1-D and 2-D systolic architectures are analyzed in terms of frequency, gate count, total power. The design is synthesized using CADENCE RTL compiler with TSMC 90nm standard cell library. The operating frequency of 1-D design is 323.20 MHz and 2-D design is 166.67 MHz and the gate count for 1-D is around 5k and for 2-D is around 21k gates and these designs can treat up to 41 Motion Vectors.
With the advent of cloud computing and its technologies, including data deduplication, more freedom are offered to the users in terms of cloud storage, processing power and efficiency, and data accessibility. The digi...
详细信息
With the advent of cloud computing and its technologies, including data deduplication, more freedom are offered to the users in terms of cloud storage, processing power and efficiency, and data accessibility. The digital data has attained unexceptional growth due to the common use of internet and digital devices giving rise to Big Data problem world wise. These huge volumes of data need some practical platforms for the storage, processing and availability and cloud technology offers all the potentials to fulfil these requirements. Data deduplication is referred to as a strategy offered to cloud storage providers (CSPs) to eliminate the duplicate data and keep only a single unique copy of it for storage space saving purpose to condense Big Data issues. But these benefits also come with data security and privacy issues associated with the cloud technology since the data owner looses the physical control of its data once uploaded in the cloud storage and the CSP gains a complete ownership of the data. In this paper, assuming that the CSP is semi-honest (i.e. Honest but curious and cannot be completely trusted), a proof of retrievability (POR) and a proof of ownership (POW) are proposed for video deduplication in cloud storage environments. The POW protocol is meant to be used by the CSP to authenticate the true owner of the data video before releasing it whereas the POR protocol is meant to allow the user to check that his/her data video stored in the cloud is secured against any malicious user or the semi-honest CSP. These schemes are proposed as complement to our earlier proposed scheme for securing the video deduplication in the cloud storage through the H.264 compression algorithm. Some experimental results are provided, showing the effectiveness of our proposed POR and POW protocols.
Progress of optical information processing is driven by two sources : optoelectronic technology benefits from the development of compound semiconductor circuit technology, and the continued growth of electronic chip s...
详细信息
Progress of optical information processing is driven by two sources : optoelectronic technology benefits from the development of compound semiconductor circuit technology, and the continued growth of electronic chip speed and density increases the need for new interconnect technologies. We advocate the use of these advances for solving complex law level image processing problems using dedicated optoelectronic processor arrays. These massively parallel, special purpose machines would not be conceivable without a suitable combination of optics with high density hybrid (digital and analog, electronic and optical) smart pixels. One optical operation of particular interest here is the generation of random number arrays for the implementation of stochastic algorithms. We describe the general structure of optoelectronic "smart retinae" that combine laser speckle, optical convolution and VLSI circuits to implement parallel simulated annealing. We report on our present results that have demonstrated one simple case of such an operation performing parallel simulated annealing at video rate.
暂无评论