The data collected by sensor networks often contain sensitive information and care must be taken to prevent that from being leaked malicious third parties, e.g., eavesdroppers. Under both the Neyman–Pearson Bayesian frameworks, we investigate strategy of defending against an informed greedy eavesdropper who has access all sensors' outputs via imperfect communication channels. Meanwhile, legitimate user, fusion center, is guaranteed achieve its desired detection performance. framework,...

This paper considers asymptotic perfect secrecy and estimation in distributed for large sensor networks under threat of an eavesdropper, which has access to all outputs. To measure secrecy, we compare the performance at fusion center eavesdropper terms their respective Fisher Information. We analyze Information ratio between derive maximum achievable when channels sensors are noisy binary symmetric channels. Furthermore, noiseless channels, show that can be made arbitrarily by careful design...

We address the optimal quantizer design problem for distributed Bayesian parameter estimation with one-bit quantization at local sensors. A performance limit obtained any estimator a known prior is adopted as guidance design. Aided by limit, and set of noisy observation models that achieve are derived. Further, when may not be achievable some applications, we develop near-optimal which consists dithered noise single threshold quantizer. In scenario where Gaussian signal-to-noise ratio...

This paper examines the secrecy in distributed detection under threat of a global eavesdropper (Eve) which has access to all sensors decisions. To measure secrecy, we compare performance at fusion center (FC) and Eve terms their respective Kullback-Leibler Divergence (KLD). When channels between FC are noiseless noisy, show that KLD ratio can be made arbitrarily large, provided log-likelihood local is unbounded. As result, perfect achieved asymptotically by making small with almost 0...

When deploying segmentation models in practice, it is critical to evaluate their behaviors varied and complex scenes. Different from the previous evaluation paradigms only consideration of global attribute variations (e.g. adverse weather), we investigate both local for robustness evaluation. To achieve this, construct a mask-preserved editing pipeline edit visual attributes real images with precise control structural information. Therefore, original labels can be reused edited images. Using...

In this paper, a performance limit is derived for distributed Bayesian parameter estimation problem in sensor networks where the prior probability density function of known. The observations are assumed conditionally independent and identically given to be estimated, sensors employ identical quantizers. established terms best possible asymptotic that scheme can achieve all observation models. This obtained by deriving optimal probabilistic quantizer under ideal setting, observe directly...

In this paper, a distributed detection model is introduced for m-ary hypotheses testing where the local sensors quantize their decisions to messages with alphabet size of D and number random following Poisson distribution. This can be applied wide variety problems including homogenous heterogeneous networks, robust under security attacks, sensor failure mode analysis. As an illustrative example, proposed Cognitive Radio network performance strategies regarding Byzantine attacks are...