Learning Recognition and Segmentation Using the Cresceptron

This paper presents a framework called Cresceptron for view-based learning, recognition and segmentation. Specifically, it recognizes and segments image patterns that are similar to those learned, using a stochastic distortion model and view-based interpolation, allowing other view points that are moderately different from those used in learning. The learning phase is interactive. The user trains the system using a collection of training images. For each training image, the user manually draws a polygon outlining the region of interest and types in the label of its class. Then, from the directional edges of each of the segmented regions, the Cresceptron uses a hierarchical self-organization scheme to grow a sparsely connected network automatically, adaptively and incrementally during the learning phase. At each level, the system detects new image structures that need to be learned and assigns a new neural plane for each new feature. The network grows by creating new nodes and connections which memorize the new image structures and their context as they are detected. Thus, the structure of the network is a function of the training exemplars. The Cresceptron incorporates both individual learning and class learning; with the former, each training example is treated as a different individual while with the latter, each example is a sample of a class. In the performance phase, segmentation and recognition are tightly coupled. No foreground extraction is necessary, which is achieved by backtracking the response of the network down the hierarchy to the image parts contributing to recognition. Several stochastic shape distortion models are analyzed to show why multilevel matching such as that in the Cresceptron can deal with more general stochastic distortions that a single-level matching scheme cannot. The system is demonstrated using images from broadcast television and other video segments to learn faces and other objects, and then later to locate and to recognize similar, but possibly distorted, views of the same objects.     

Some aspects of fluidized bed combustion of paddy husk

Some agricultural wastes—for example, bagasse, paddy husk, etc.—are frequently used as fuels, paddy husk appearing, in particular, to be quite suitable for fluidized bed combustion. The conventional method of combustion of paddy husk in grate-type furnaces is slow and inefficient.

This paper reports certain aspects of the fluidized bed combustion of paddy husk. Fluidized bed combustion was carried out by feeding husk in a bed of sand particles. The unexpanded bed height was 10 cm and the size of the sand particles, 351–420 μm. The superficial velocity of the ambient fluidization air through the bed ranged from 11·1 to 22·2 m/min.

A combustion intensity of about 530 kg/h/m² of distributor area could be achieved. This is about 7·5 times higher than the maximum combustion intensity possible in a grate-type furnace per unit grate area. The efficiency of combustion, which ranged from 81 to 98 per cent was found to increase with the air flow rate. There was significant carry-over of inert sand particles from the bed under conditions of high air flow rate.

Combustion intensity increased as the bed height rose from 10 cm to 15 cm, but increased sand entrainment also occurred.     

Comparative sorption equilibrium studies of toxic phenols on flyash and impregnated flyash

In the present study, the sorption of toxic phenols, which include phenol, o-cresol, m-cresol, p-cresol, o-nitrophenol, m-nitrophenol and p-nitrophenol, has been investigated. The influences of various factors, such as particle size, impregnation of flyash (IFA), pH and temperature on the sorption capacity have been studied. Equilibrium modelling has been carried out using Langmuir and Freundlich isotherm equations and constants have been calculated under different conditions. Thermodynamic studies have also been carried out and values of standard free energy (ΔG°), enthalpy (ΔH°) and entropy (ΔS°) calculated.     

Comparative sorption kinetic studies of phenolic compounds on fly ash and impregnated fly ash

The factors affecting the rate processes involved in the removal of phenolic compounds, e.g. initial phenol concentration, particle size, impregnation of fly ash (IFA), pH and temperature have been studied. The removal rate of phenols varied in the order p-nitrophenol m-nitrophenol > o-nitrophenol > p-cresol > phenol > m-cresol > o-cresol. The process followed first order rate kinetics. The sorption data generally fit the Lagergren equation and the intraparticle diffusion rate equation from which adsorption rate constants, diffusion rate constants and diffusion coefficients were determined. Intraparticle diffusion was found to be the rate-limiting step. These kinetic parameters were compared for various phenols under different conditions using fly ash (FA) and impregnated fly ash (IFA).     

The use of learning algorithms in telephone traffic routing—A methodology

A theoretical as well as conceptual framework for the use of learning algorithms in telephone traffic routing is given. The approach is distinctly different from the mathematical programming methods generally used in such cases. Learning algorithms at the network nodes update their strategies for routing traffic on the basis of success or failure in completing calls. The entire system is described as a Markov process and different learning schemes are shown to lead to different flow patterns in the steady state.     

A Cassava Starch–Chitosan Edible Coating Enriched with <Emphasis Type="Italic">Lippia sidoides</Emphasis> Cham. Essential Oil and Pomegranate Peel Extract for Preservation of Italian Tomatoes (<Emphasis Type="Italic">Lycopersicon esculentum</Emphasis> Mill.) Stored at Room Temperature

The effect of edible cassava starch–chitosan coatings incorporated with rosemary pepper (Lippia sidoides Cham.) essential oil and pomegranate peel extract on the shelf-life of tomatoes during storage at 25 °C for 12 days was investigated. Sixteen formulations, containing 10 g L^?1 cassava starch and various concentrations of chitosan (5, 10, 20, 30 g L^?1), essential oil (0, 2.5, 5, 10 mL L^?1) and pomegranate peel extract (0, 5, 10, 20 mL L^?1) were prepared and applied to tomatoes. Physical–chemical and microbiological analyses were performed on days 1, 4, 8 and 12. Most of the coatings delayed the ripening of tomatoes, lowering the total soluble solids (38?44 g sucrose kg^?1) and weight loss (93?128 g kg^?1) and maintaining constant firmness compared to the uncoated tomatoes (45 g sucrose kg^?1, 175 g kg^?1) at 12 days of storage. Conversely, except red intensity (a*), which was higher for the uncoated samples, the colour parameters (L*, b*) of the coated and control tomatoes were similar at the end of storage. Uncoated and coated tomatoes showed no contamination during storage. The coatings showed potential to maintain the quality of tomatoes during storage at 25 °C for 12 days. In this context, tomatoes coated with the formulation comprising 10 g L^?1 cassava starch, 10 g L^?1 chitosan, 10 mL L^?1 essential oil and 20 mL L^?1 pomegranate peel extract showed the lowest weight loss and reduced total soluble solids content compared with uncoated ones.     

Genesis of enhanced photoactivity of CdS/Nix nanocomposites for visible-light-driven splitting of water

A series of CdS/Ni_x nanocomposite photocatalysts, containing ca. 0.6–15 wt% Ni, were synthesized using a one-step hydrothermal method and characterized for their crystallographic, morphological, interfacial, and optical properties. Rietveld refinement of powder XRD data revealed the coexistence of wurtzite (hexagonal) and zinc blende (cubic) phases of CdS in ratios dependent on Ni content. Only a fraction of Ni existed as a secondary phase of NiS while the majority occupied the lattice positions of hexagonal CdS. Whereas up to 10-fold enhancement in H₂ evolution compared with pure CdS was observed for samples containing ∼1.5–4.5 wt% Ni, samples with smaller or larger Ni content displayed poor activity for visible-light-induced splitting of water in presence of sulfide–sulphite ions as sacrificial electron donors. In contrary to recent findings, our study reveals that the enhanced CdS photoactivity is not a result of charge transfer between p-type NiS and n-type CdS, Ni-induced visible-region absorbance, or the coating of CdS particles by non-photoactive NiS. Instead, the preparation-dependent hexagonal/cubic CdS phase boundaries and particle morphology may play a crucial role. Additionally, certain Ni-doping-induced sub-bandgap shallow energy levels contribute to charge carrier separation.