Most current algorithms for blind steganalysis of images are based on a two-stages approach: First, features are extracted in order to reduce dimensionality and to highlight potential manipulations; second, a classifier trained on pairs of clean and stego images finds a decision rule for these features to detect stego images. Thereby, vector components might vary significantly in their values, hence normalization of the feature vectors is crucial. Furthermore, most classifiers contain free parameters, and an automatic model selection step has to be carried out for adapting these parameters. However, the commonly used cross-validation destroys some information needed by the classifier because of the arbitrary splitting of image pairs (stego and clean version) in the training set. In this paper, we propose simple modifications of normalization and for standard cross-validation. In our experiments, we show that these methods lead to a significant improvement of the standard blind steganalyzer of Lyu and Farid

The heat capacity of the adsorbate−adsorbent system is a crucial physical property required for modeling adsorption cycles. It is specific to the considered adsorptive gas and to the corresponding adsorbent. In previous work, it has been shown that this heat capacity can be estimated using only the properties of the gaseous phase, the adsorption equilibria, and the isosteric heats of adsorption. As these quantities are required in any case for simulations, no additional effort arises. In this paper, the separated adsorbed phase heat capacities are computed at higher accuracies by employing reversible thermodynamic paths, which traverse equilibrium states only. As a result, the thermodynamic consistency of the model improves, since the residuals of the energy and entropy balances are significantly smaller. Results are given for the water−zeolite 13X adsorption pair.

A novel thermodynamic cycle for adsorption heat pumps and chillers is presented. It shows a significant improvement of the internal heat recovery between the adsorption and the desorption half cycle. A stratified thermal storage, which allows for a temperature-based extraction and insertion of storage fluid, is hydraulically coupled with a single adsorber. The benefit is an increased efficiency by reusing the released heat of adsorption for regeneration of the adsorber and by rendering possible low driving temperature differences. For investigating the second law of this cycle, a dynamic model is employed. The transient behavior of the system and the respective losses because of driving temperature differences at the heat exchangers and losses due to mixing within the storage and to the surroundings are depicted in this one-dimensional model. The model is suitable both for analyzing this advanced cycle as well as for comparisons with other cycles.

A thermodynamically consistent modeling of adsorption equilibrium data is essential in the modeling of adsorption heat pump cycles. Especially when experimental data are scarce, the approach taken by Dubinin and co-workers to map all adsorption data onto a single “characteristic curve” can be very useful. Different choices for obtaining this curve for modeling purposes are discussed and analyzed here using adsorption equilibrium data of water on zeolite Li-LSX as an example. It is shown that the fit function proposed by Dubinin and Astakhov results in a poor fit to the data. The use of arbitrary fit functions, however, leads to an overfitting of the data, as is exemplified through a cross-validation analysis. Gaussian process regression (GPR) is discussed as an alternative way to compute the “characteristic curve” and is shown to result in a good fit to the data while, by design, avoiding overfitting.

Abstract. We compare two image bases with respect to their capabilities for image modeling and steganalysis. The first basis consists of wavelets, the second is a Laplacian pyramid. Both bases are used to decompose the image into subbands where the local dependency structure is modeled with a linear Bayesian estimator. Similar to existing approaches, the image model is used to predict coefficient values from their neighborhoods, and the final classification step uses statistical descriptors of the residual. Our findings are counter-intuitive on first sight: Although Laplacian pyramids have better image modeling capabilities than wavelets, steganalysis based on wavelets is much more successful. We present a number of experiments that suggest possible explanations for this result.