This paper presents a framework of on-demand color calibration system to track pedestrians across nonoverlapping fields of fixed camera view. The proposed system is designed based on the machine-to-machine (M2M) approach, and exchanges color information of multiple fixed cameras autonomously. The fixed cameras are assumed to be pointing in different directions and have nonoverlapping fields of view. The color information is extracted when a pedestrian vacates from the field of view of a camera, and the information will be sent to an adjacent camera and used for color calibration automatically. The automatic calibration uses identical objects whose original colors are the same but captured different color in views of different cameras. Using this color information of the identical objects, the color calibration matrix is calculated by the camera receiving the color information. Experiment results indicate that our proposed system effectively addresses the seamless pedestrian tracking in nonoverlapping areas.

Crumpled sheets of paper tend to exhibit specific and complex structure, which is usually described as ridge network by physicists. Existing literature has showed that it is difficult to automate ridge network detection in crumpled paper because of its complex structure. In this paper, we attempt to develop an automatic detection process in terms of our proposed density criterion. We model the ridge network as a weighted graph, where the nodes indicate the intersections of ridges and the edges are the straightened ridges detected in crumpled paper. We construct the weighted graph by first detecting the nodes and then determining the edge weight using the ridge responses. Next, we formulate a graph density criterion to evaluate the detected ridge network. Finally, we propose an edge linking method to construct the graph by maximizing the proposed density criterion. Our experimental results show that, with the density criterion, our proposed node detection together with the edge line linking method could effectively automate the ridge network detection.

Boltzmann machines can be regarded as Markov random fields. For binary cases, they are equivalent to the Ising spin model in statistical mechanics. Learning systems in Boltzmann machines are one of the NP-hard problems. Thus, in general we have to use approximate methods to construct practical learning algorithms in this context. In this letter, we propose new and practical learning algorithms for Boltzmann machines by using the belief propagation algorithm and the linear response approximation, which are often referred as advanced mean field methods. Finally, we show the validity of our algorithm using numerical experiments.

Abstract We show that each of and proves -Det and that neither nor can be dropped. We also show that neither nor proves -Det. Moreover, we prove that none of and is provable in

Abstract In this paper, we show within RCA0 that weak König's lemma is necessary and sufficient to prove that any (separable) compact group has a Haar measure. Within WKL0, a Haar measure is constructed by a non-standard method based on a fact that every countable non-standard model of WKL0 has a proper initial part isomorphic to itself [10].

The Markov random field (MRF) model, which is a method in the image restoration, is reformulated, based on the variational principle of the difference minimum between the original and damaged image datas, under the constraints for the area of black faces, the boundary length between the different darkness and the number of corners in the edge. Moreover, a new method is proposed for the estimation of the areas of black faces, the boundary length between the different darkness and the number of corners in the edge from only the damaged image.