Object Bank: An Object-Level Image Representation for High-Level Visual Recognition

It is a remarkable fact that images are related to objects constituting them. In this paper, we propose to represent images by using objects appearing in them. We introduce the novel concept of object bank (OB), a high-level image representation encoding object appearance and spatial location information in images. OB represents an image based on its response to a large number of pre-trained object detectors, or ‘object filters’, blind to the testing dataset and visual recognition task. Our OB representation demonstrates promising potential in high level image recognition tasks. It significantly outperforms traditional low level image representations in image classification on various benchmark image datasets by using simple, off-the-shelf classification algorithms such as linear SVM and logistic regression. In this paper, we analyze OB in detail, explaining our design choice of OB for achieving its best potential on different types of datasets. We demonstrate that object bank is a high level representation, from which we can easily discover semantic information of unknown images. We provide guidelines for effectively applying OB to high level image recognition tasks where it could be easily compressed for efficient computation in practice and is very robust to various classifiers.     

A memetic algorithm for the capacitated m-ring-star problem

The Capacitated m-Ring-Star Problem (CmRSP) models a network topology design problem in the telecommunications industry. In this paper, we propose to solve this problem using a memetic algorithm that includes a crossover operation, a mutation operation, a local search involving three neighborhood operators, and a population selection strategy that maintains population diversity. Our approach generates the best known solutions for 131 out of 138 benchmark instances, improving on the previous best solutions for 24 of them, and exhibits more advantages on large benchmark instances when compared with the best existing approach. Additionally, all existing algorithms for this problem in literature assume that the underlying graph of the problem instance satisfies the triangle inequality rule; our approach does not require this assumption. We also generated a new set of 36 larger test instances based on a digital data service network price structure to serve as a new benchmark data set for future researchers.     

Domain ontology graph model and its application in Chinese text classification

This paper proposes an ontology learning method which is used to generate a graphical ontology structure called ontology graph. The ontology graph defines the ontology and knowledge conceptualization model, and the ontology learning process defines the method of semiautomatic learning and generates ontology graphs from Chinese texts of different domains, the so-called domain ontology graph (DOG). Meanwhile, we also define two other ontological operations—document ontology graph generation and ontology graph-based text classification, which can be carried out with the generated DOG. This research focuses on Chinese text data, and furthermore, we conduct two experiments: the DOG generation and ontology graph-based text classification, with Chinese texts as the experimental data. The first experiment generates ten DOGs as the ontology graph instances to represent ten different domains of knowledge. The generated DOGs are then further used for the second experiment to provide performance evaluation. The ontology graph-based approach is able to achieve high text classification accuracy (with 92.3 % in f-measure) over other text classification approaches (such as 86.8 % in f-measure for tf–idf approach). The better performance in the comparative experiments reveals that the proposed ontology graph knowledge model, the ontology learning and generation process, and the ontological operations are feasible and effective.     

Modeling of rotational molding process: Multi‐layer slip‐flow model,phase‐change,and warpage

A new multilayer slip‐flow model has been developed to simplify and to overcome current numerical difficulties of two‐dimensional model in predicting the internal air temperature inside a mold during a rotational molding process. The proposed methodology considers a macroscopic “layer‐by‐layer” deposition of a heating polymer bed onto the inner mold surface. A semi‐implicit approach is introduced and applied to compute the complex thermal interactions between the internal air and its surroundings. In the model, the lumped‐parameter system and the coincident node technique are incorporated with the Galerkin finite element model to address the internal air and the deposition of molten polymer beds, respectively. The simple phase‐change algorithm has been proposed to improve the computational cost, numerical nonlinearity, and predicted results. The thermal aspects of the inherent warpage are explored to study its correlation to the weak apparent crystallization‐induced plateau in the temperature profile of the internal air, as in practice. The overall predicted results are in favor with the available experimental data for rotomolded parts of cross‐sectional thicknesses up to 12 mm. POLYM. ENG. SCI. 46:960–969, 2006. © 2006 Society of Plastics Engineers     

A novel layered Li [Li0.12NizMg0.32−zMn0.56]O2 cathode material for lithium-ion batteries

Layered Li[Li_0.12Ni_zMg_0.32−zMn_0.56]O₂ oxide cathodes containing lithium atoms in the transition metal layers were synthesized and characterized using X-ray diffraction (XRD), galvanostatic cycling, and differential scanning calorimetry (DSC). The Li[Li_0.12Ni_zMg_0.32−zMn_0.56]O₂ cathodes deliver a specific discharge capacity of about 190 mAh/g at room temperature and 236 mAh/g at 55 °C when cycled between 2.7 and 4.6 V versus Li/Li⁺. Excellent capacity retention and smooth potential profiles at room and elevated temperatures over extended cycles suggest that this material does not convert into a spinel structure.     

Characteristics of selective adsorption using D-phenylalanine imprinted terpolymer beads

A D-Phenylalanine (Phe) imprinted terpolymer, Polyacrylonitrile-Poly(acrylic acid)-Poly(acryl amide) (Poly(AN-AA-AAm)) bead was prepared by the wet-phase inversion method. Acrylamide (AAm) and acrylic acid (AA) were used as the functional monomer and acrylonitrile (AN) was used as a physical cross linker. The characteristics of selective adsorption by the D-Phe imprinted terpolymer beads were investigated at high concentrations of Phe racemate solution, 1 g Phe/L, and 10 g Phe/L. The adsorption selectivity of the D-Phe imprinted terpolymer beads prepared by anin-situ implanting method reached 0.82 and 0.8 at 1.0 g and 10 g Phe/L racemate solution, respectively, and almost all of the adsorbed D-Phe and about 43% of the adsorbed L-Phe were desorbed by 4% acetic acid. The uptake capacities of the terpolymer beads were maintained for several repeated batches.     

Preparation and characterization of syndiotactic polystyrene/ethylene‐propylene copolymer blends

The reactive compatibilization of syndiotactic polystyrene (sPS)/oxazoline‐styrene copolymer (RPS)/maleic anhydride grafted ethylene‐propylene copolymer (EPR‐MA) blends is investigated in this study. First, the miscibility of sPS/RPS blends is examined by thermal analysis. The cold crystallization peak (T_cc) moved toward higher temperature with increased PRS, and, concerning enthalpy relaxation behaviors, only a single enthalpy relation peak was found in all aged samples. These results indicate that the sPS/RPS blend is miscible along the various compositions and RPS can be used in the reactive compatibilization of sPS/RPS/EPR‐MA blends. The reactive compatibilized sPS/RPS/EPR‐MA blends showed finer morphology than sPS/EPR‐MA physical blends and higher storage modulus (G') and complex viscosity (η*) when RPS contents were increased. Moreover, the impact strength of sPS/RPS/EPR‐MA increased significantly compared to sPS/EPR‐MA blend, and SEM micrographs after impact testing show that the sPS/RPS/EPR‐MA blend has better adhesion between the sPS matrix and the dispersed EPR‐MA phase. © 2002 Wiley Periodicals, Inc. J Appl Polym Sci 85: 2084–2091, 2002     

Application of a fibre-reactive chitosan derivative to cotton fabric as a zero-salt dyeing auxiliary

Cotton fabric has been treated with a fibre-reactive chitosan derivative containing quaternary ammonium groups, O -acrylamidomethyl- N -[(2-hydroxy-3-trimethylammonium)propyl]chitosan chloride (NMA-HTCC). Cotton treated with NMA-HTCC has been dyed with direct and reactive dyes without the addition of salt. The colour yield was higher than that on untreated cotton, despite the addition of a large amount salt in the latter case. After dyeing, cotton treated with NMA-HTCC gave better wash fastness than the untreated cotton. The light fastness was however inferior to that on untreated cotton. The antimicrobial activity of cotton treated with NMA-HTCC against Staphylococcus aureus was considerably lower after dyeing, probably due to the antimicrobial effect of the cationic group on NMA-HTCC being blocked by its combination with the anionic dye.     

Enhanced surface morphologies of screen-printed carbon nanotube films by heat treatment and their field-emission properties

A heating process for obtaining free-standing carbon nanotube emitters is presented with the aim of improving field-emission properties from the screen-printed multiwalled carbon nanotube (MWCNT) films. Using an atmosphere with an optimum combination of nitrogen and air for heat treatment of CNT films, the CNT emitters can be made to protrude from the surface. This allows for a high emission current and the formation of very uniform emission sites without special surface treatment. The morphological change of the CNT film by this technique has eliminated additional processing steps, such as surface treatment which may result in secondary contamination and damage to the film. Despite its simplicity the process provides a high reproducibility in emission current density which makes the films suitable for practical applications.     

Thermodynamic Analysis of Chemical Vapor Deposition of BN + AIN Composite Coatings

Thermodynamic calculations were performed using a modified solgasmix-pv computer program in order to study the feasibility of codepositing boron nitride (BN) plus aluminum nitride (AIN) by chemical vapor deposition. Reactants considered were AICl₃, BCl₃ or B₂H₆, NH₃, and H₂. Deposition diagrams were generated for the BCl₃-AICl₃-NH₃ system over a range of processing conditions such as temperature, total system pressure, and reagent concentrations. Codeposition of BN + AIN was predicted by the calculations for temperatures in the range of 900 to 1700 K and pressures of 10.13 to 101.3 kPa. The predicted deposition efficiency at equilibrium was much higher for BN than for AlN at most reagent compositions. The AlN deposition efficiency increased with decreasing temperature and decreasing BCl₃ content, with increasing NH₃ content, or with the addition of H₂. Aluminum chlorides were found to be the dominant gaseous species.