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     

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.     

Experimental Simulation of the Reactor Section of Fluid Cokers: Comparison of FCC and Fluid Coke Particles

The hydrodynamics of fluid cokers were studied in a pressurized fully cylindrical cold model of diameter 483 mm, geometrically and dynamically scaled down by a factor of ～20 from commercial units. Differential pressure fluctuations, voidage distributions, solids momentum flux distributions and steady state gas mixing behaviour in the reactor section are compared for the same operating conditions with two kinds of particles, FCC and fluid coke. The voidage distributions and core‐annular flow structures in the reactor section were similar enough that either FCC or fluid coke particles can be used for cold modelling of fluid cokers.     

Preparation of pH‐sensitive poly(vinyl alcohol‐g‐methacrylic acid) and poly(vinyl alcohol‐g‐acrylic acid) hydrogels by gamma ray irradiation and their insulin release behavior

A series of pH‐responsive hydrogels were studied as potential drug carriers for the protection of insulin from the acidic environment of the stomach before releasing in the small intestine. Hydrogels based on poly(vinyl alcohol) networks grafted with acrylic acid or methacrylic acid were prepared by a two‐step process. Poly(vinyl alcohol) hydrogels were prepared by gamma ray irradiation (50 kGy) and then followed by grafting either acrylic acid or methacrylic acid onto these poly(vinyl alcohol) hydrogels with subsequent irradiation (5–20 kGy). These graft hydrogels showed pH‐sensitive swelling behavior and were used as carriers for the controlled release of insulin. The in vitro release of insulin was observed for the insulin‐loaded hydrogels in a simulated intestinal fluid (pH 6.8) but not in a simulated gastric fluid (pH 1.2). The release behavior of insulin in vivo in a rat model confirmed the effectiveness of the oral delivery of insulin to control the level of glucose. © 2003 Wiley Periodicals, Inc. J Appl Polym Sci 91: 636–643, 2004     

Non-Isothermal crystallization of isotactic polypropylene in dotriacontane. II: Effects of dilution,cooling rate,and nucleating agent addition on growth rate

Non-isothermal crystallization growth rates of nucleated and non-nucleated isotactic polypropylene (iPP) in dotriacontane were determined experimentally by thermal optical microscopy. Adipic acid was used as the nucleating agent. The non-isothermal growth rates of the nucleated and non-nucleated systems were compared with experimentally determined isothermal growth rates. The Lauritzen and Hoffman growth rate equation, originally developed for isothermal crystallization, was modified to describe the non-isothermal growth. The modified Lauritzen-Hoffman equation was used to predict isothermal growth rates from non-isothermal crystallization for the nucleated and non-nucleated polymer-diluent mixtures. This study is the first to deal with polymer-diluent-nucleating agent systems vital to membrane production.     

Wetting properties of polystyrene ionomers treated with plasma source ion implantation

The wetting properties of polystyrene‐based ionomers treated with plasma source ion implantation (PSII) were investigated by the measurement of water contact angles. When sulfonated ionomers were aged for a few days, the hydrophobic recovery for the ionomers became much slower than that for the nonionic polymers. However, when the samples were aged over 20 days, the water contact angle of the ionomers converged with that of the nonionic polymer. Thus, it was concluded that the ionic interaction between the ionic groups and the presence of ionic groups together resulted in the slow hydrophobic recovery and that the aging effect was significant for the ionomers. For the methacrylate ionomer of low ion content, on the other hand, it was found that the PSII treatment produced only a small change in hydrophobic recovery behavior. Thus, it was suggested that the low ionic content coupled with the small size of the ionic unit might cause changes only of a very insignificant degree in hydrophobic recovery behavior. © 2002 Wiley Periodicals, Inc. J Appl Polym Sci 83: 2500–2504, 2002