Attribute Reduction in the Bayesian Version of Variable Precision Rough Set Model

The article presents a parametric Bayesian extension of the rough set model, where the set approximations are defined by using the prior probability as a reference. It is shown that the quality of the Bayesian rough set models can be evaluated using probabilistic gain function which is suitable for easy computation of attribute reducts. It leads to the Bayesian style criteria for the attribute reduction within the rough set framework.     

Deterministic broadcasting in ad hoc radio networks

We consider the problem of distributed deterministic broadcasting in radio networks of unknown topology and size. The network is synchronous. If a node u can be reached from two nodes which send messages in the same round, none of the messages is received by u. Such messages block each other and node u either hears the noise of interference of messages, enabling it to detect a collision, or does not hear anything at all, depending on the model. We assume that nodes know neither the topology nor the size of the network, nor even their immediate neighborhood. The initial knowledge of every node is limited to its own label. Such networks are called ad hoc multi-hop networks. We study the time of deterministic broadcasting under this scenario. For the model without collision detection, we develop a linear-time broadcasting algorithm for symmetric graphs, which is optimal, and an algorithm for arbitrary n-node graphs, working in time . Next we show that broadcasting with acknowledgement is not possible in this model at all. For the model with collision detection, we develop efficient algorithms for broadcasting and for acknowledged broadcasting in strongly connected graphs. Received: January 2000 / Accepted: June 2001     

Subadditivity of Entropy for Stochastic Matrices

The entropy of an irreducible stochastic matrix measures its mixing properties. We show that this quantity is subadditive and strongly subadditive.     

Hiding information in a Stream Control Transmission Protocol

The STCP (Stream Control Transmission Protocol) is a candidate for a new transport layer protocol that may replace the TCP (Transmission Control Protocol) and the UDP (User Datagram Protocol) protocols in future IP networks. Currently, the SCTP is implemented in, or can be added to, many popular operating systems (Windows, BSD, Linux, HP-UX or Sun Solaris). This paper identifies and presents the most likely “places” where hidden information can be exchanged using an SCTP. The paper focuses mostly on proposing new steganographic methods that can be applied to an SCTP and that can utilise new, characteristic SCTP features, such as multi-homing and multi-streaming. Moreover, for each method, the countermeasure is covered. When used with malicious intent, a method may pose a threat to network security. Knowledge about potential SCTP steganographic methods may be used as a supplement to RFC5062, which describes security attacks in an SCTP protocol. Presented in this paper is a complete analysis of information hiding in an SCTP, and this analysis can be treated as a “guide” when developing steganalysis (detection) tools.     

Multi‐Layered Automultiscopic Displays

Our hybrid display model combines multiple automultiscopic elements volumetrically to support horizontal and vertical parallax at a larger depth of field and better accommodation cues compared to single layer elements. In this paper, we introduce a framework to analyze the bandwidth of such display devices. Based on this analysis, we show that multiple layers can achieve a wider depth of field using less bandwidth compared to single layer displays. We present a simple algorithm to distribute an input light field to multiple layers, and devise an efficient ray tracing algorithm for synthetic scenes. We demonstrate the effectiveness of our approach by both software simulation and two corresponding hardware prototypes.     

Design of a device for precision shaping of grinding wheel macro- and micro-geometry

The selected modifications to the construction of grinding wheels were described which facilitate an increase in the material removal rate (grinding wheels with conic chamfer and grinding wheels with microdiscontinuities on the active surface). Using these background details, a suggested thesis was put forward regarding the need to develop a device which will allow for the shaping of the macrogeometry of the grinding wheel (cylindrical and conical surfaces) and the microdiscontinuities within the dressing operation simultaneously. The device was presented and prepared in two functional variants (horizontal and vertical mounting of the motor), then a prototype was described. An example of the grinding wheel active surface, shaped by using this device, was also presented. The theoretical analysis and experimental verification performed determine that the error of shaping the conic chamfer angle within the range of 0–1.5°, using the developed device, is approximately ±3%.     

Enhanced representation and multi-task learning for image annotation

In this paper we propose a novel biased random sampling strategy for image representation in Bag-of-Words models. We evaluate its impact on the feature properties and the ranking quality for a set of semantic concepts and show that it improves performance of classifiers in image annotation tasks and increases the correlation between kernels and labels. As second contribution we propose a method called Output Kernel Multi-Task Learning (MTL) to improve ranking performance by transfer information between classes. The main advantages of output kernel MTL are that it permits asymmetric information transfer between tasks and scales to training sets of several thousand images. We give a theoretical interpretation of the method and show that the learned contributions of source tasks to target tasks are semantically consistent. Both strategies are evaluated on the ImageCLEF PhotoAnnotation dataset.Our best visual result which used the MTL method was ranked first according to mean Average Precision (mAP) within the purely visual submissions in the ImageCLEF 2011 PhotoAnnotation Challenge. Our multi-modal submission achieved the first rank by mAP among all submissions in the same competition.     

Dispersion‐based Color Projection using Masked Prisms

We present a method for projecting arbitrary color images using a white light source and an optical device with no colored components — consisting solely of one or two prisms and two transparent masks. When illuminated, the first mask creates structured white light that is then dispersed in the prism and attenuated by the second mask to create the color projection. We derive analytical expressions for the mask parameters from the physical components and validate our approach both in simulation and also demonstrate it on a wide variety of images using two different physical setups (one consisting of two inexpensive triangular prisms, and the other using a single rhombic prism). Furthermore, we show that optimizing the masks simultaneously enables obfuscating the image content, and provides a tradeoff between increased light throughput (by up to a factor of three) and maximum color saturation.     

Contextual reliability discounting in welding process diagnostic based on DSmT

The monitoring and diagnosing of the welding process is an important task in assuring the quality of welded joints. Observation of such a dynamic process is often performed by measuring various signals of different modality. The main problem is how to use several of the possible signals simultaneously. Many techniques that can deal with this kind of task were elaborated and introduced in practical application. One of these is decision fusion, with a group of methods based on belief function theories: Dempster–Shafer and Dezert–Smarandache (DSmT). In this paper, extended fusion scheme for assessment of welding process condition, with contextual discounting of evidence sources reliability in DSmT framework, is presented. Discounting operation was used to distinguish the objective ability of each evidence source, to give a correct assessment of process condition. To validate this method, several series of welded joints representing different process conditions that result in various joint quality were made. On the basis of performed research, it was proven that contextual reliability discounting leads to increase of the condition recognition accuracy.