Distributed intelligence for multi-camera visual surveillance

Latest advances in hardware technology and state of the art of computer vision and artificial intelligence research can be employed to develop autonomous and distributed monitoring systems. The paper proposes a multi-agent architecture for the understanding of scene dynamics merging the information streamed by multiple cameras. A typical application would be the monitoring of a secure site, or any visual surveillance application deploying a network of cameras. Modular software (the agents) within such architecture controls the different components of the system and incrementally builds a model of the scene by merging the information gathered over extended periods of time. The role of distributed artificial intelligence composed of separate and autonomous modules is justified by the need for scalable designs capable of co-operating to infer an optimal interpretation of the scene. Decentralizing intelligence means creating more robust and reliable sources of interpretation, but also allows easy maintenance and updating of the system. Results are presented to support the choice of a distributed architecture, and to prove that scene interpretation can be incrementally and efficiently built by modular software.     

Multiple costs based decision making with back-propagation neural networks

The current research investigates a single cost for cost-sensitive neural networks (CNN) for decision making. This may not be feasible for real cost-sensitive decisions which involve multiple costs. We propose to modify the existing model, the traditional back-propagation neural networks (TNN), by extending the back-propagation error equation for multiple cost decisions. In this multiple-cost extension, all costs are normalized to be in the same interval (i.e. between 0 and 1) as the error estimation generated in the TNN. A comparative analysis of accuracy dependent on three outcomes for constant costs was performed: (1) TNN and CNN with one constant cost (CNN-1C), (2) TNN and CNN with two constant costs (CNN-2C), and (3) CNN-1C and CNN-2C. A similar analysis for accuracy was also made for non-constant costs; (1) TNN and CNN with one non-constant cost (CNN-1NC), (2) TNN and CNN with two non-constant costs (CNN-2NC), and (3) CNN-1NC and CNN-2NC. Furthermore, we compared the misclassification cost for CNNs for both constant and non-constant costs (CNN-1C vs. CNN-2C and CNN-1NC vs. CNN-2NC). Our findings demonstrate that there is a competitive behavior between the accuracy and misclassification cost in the proposed CNN model. To obtain a higher accuracy and lower misclassification cost, our results suggest merging all constant cost matrices into one constant cost matrix for decision making. For multiple non-constant cost matrices, our results suggest maintaining separate matrices to enhance the accuracy and reduce the misclassification cost.     

Extracting the optimal dimensionality for local tensor discriminant analysis

Supervised dimensionality reduction with tensor representation has attracted great interest in recent years. It has been successfully applied to problems with tensor data, such as image and video recognition tasks. However, in the tensor-based methods, how to select the suitable dimensions is a very important problem. Since the number of possible dimension combinations exponentially increases with respect to the order of tensor, manually selecting the suitable dimensions becomes an impossible task in the case of high-order tensor. In this paper, we aim at solving this important problem and propose an algorithm to extract the optimal dimensionality for local tensor discriminant analysis. Experimental results on a toy example and real-world data validate the effectiveness of the proposed method.     

Improved diagnosis using enhanced fault dominance

Defect diagnosis can benefit from fault dominance relations to reduce the set of defect candidate sites. This paper presents new fault dominance collapsing operators that further reduce the set of candidates considered during the initial phase of diagnosis. In contrast to existing dominance-based methods which operate on pairs of faults, the proposed method operates on sets of faults. Fault-related entities are generated to guide the diagnosis process. The proposed collapsing operators can be used to accelerate effect-cause diagnosis. Experimental results demonstrate that the proposed method achieves a higher collapsing ratio than existing methods.     

WIPO Panels’ interpretation of the Uniform Dispute Resolution Policy (UDRP) three-prong test

The Uniform Dispute Resolution Policy (UDRP) is a unique process that resolves domain name disputes effectively and inexpensively. This paper, through an analysis of the UDRP three-prong test, revealed that even though the UDRP affords a great degree of discretion to the WIPO Panels deciding any given case, there is some consistency and predictability inherent in the UDRP process.     

Europe should stop taxing innovation

After nearly 48 years of failure to create the EU patent, language issues and the design of a centralised patent-litigation court still dominate headlines. But behind these issues there are high financial stakes and control power to play for. The recent EU Council deal on an ‘enhanced’ European patent system does not solve the above problems, and has not eliminated lingering governance issues. The risk for Europe is that a final patent agreement might be reached that does not cure the system of its major ills, and thus does not bring about any significant improvement for those that need it most: entrepreneurs and innovative companies starting out on the innovation process. The creation of an effective single EU patent requires (i) English-only post-grant translation, (ii) the end of nationally granted patents, (iii) phasing-out of the current ‘European patent’, (iv) lower fees for young innovative companies, and (v) a radical shake-up of the governance of the European Patent Office.     

Choice of a pertinent color space for color texture characterization using parametric spectral analysis

This article presents a comparison of different color spaces including RGB, IHLS and L?a*b* for color texture characterization. This comparison is based on the fusion of the independent spatial structure and color feature cues. In IHLS and L*a*b*, two channel complex color images are created from the luminance and the chrominance values. For such images, two dimensional complex multichannel linear prediction models are used to perform parametric power spectrum estimation and the structure feature cues are computed from this estimated power spectrum. Quantitative comparison of auto spectra of luminance and combined chrominance channels for different color spaces is done. This comparison is based on the degree of decorrelation between luminance and chrominance information provided by different color space transformations. Three dimensional histograms are used as color feature cues. Then, to classify color textures, Kullback-Leibler divergence based symmetric distance measures are calculated for pure color, luminance structure and chrominance structure feature cues. Individual as well as combined effect of information from all feature cues on classification results is then compared for different color spaces and different color texture data sets. The proposed color texture classification method performs better than the state of the art methods in certain cases. The L*a*b* color space gives us a better characterization of the chrominance spatial structure as well as the overall spatial structure for all of the chosen data sets. Experimental results on pixel classification of color textures are also presented and discussed.     

A fast parallel (23, 12, 7) QR decoder based on syndrome-weight determination

In this paper, a time-area efficient hardware decoder of the (23, 12, 7) quadratic residue code, or Golay code, is presented. The key feature of this proposed algorithm lies in fast determination of error positions through the analysis on the weight of syndromes without large operations in finite fields. Comparing with the algorithm using one syndrome, the proposed algorithm is more suitable for the parallel hardware design by using two syndromes. Based on an Altera Cyclone II field-programmable gate array (FPGA) device (EP2C35F484C6), the area cost and the time delay of the complete system are reduced by up to 86.4% and 22.5%, respectively. Using the Taiwan Semiconductor Manufacturing Company (TSMC) 0.18-μm complementary metal-oxide-semiconductor (CMOS) standard cell library, the proposed decoder is 91.8% smaller and 8.3% faster.     

Additive noise removal using a novel fuzzy-based filter

The paper proposes a new fuzzy-based two-step filter for restoring images corrupted with additive noise. The goal of the first step is to compute the difference between the central pixel and its neighborhood in a selected window and to compute a fuzzy membership degree for each difference value using a Gaussian membership function. Computed fuzzy membership values are appropriately utilized as weights for each pixel and then computes the weighted average representing the modified value for the current central pixel. The second step is used as an augmented step to the first one and its goal is to improve the result obtained in the first step by reducing the noise in the color component differences without destroying the fine details of the image. The experimental analysis shows that the proposed method gives better results compared to existing advanced filters for additive noise reduction. Both visual, quantitative and qualitative analysis have been done to prove the efficiency and effectiveness of the proposed method.