On the Mitchell similarity measure and its application to pattern recognition

This paper is a response to the similarity measure and pattern recognition problem of Mitchell that was published in Pattern Recognition Letters, 2003. The purpose of this paper is threefold. First, we reviewed and revised her computation for similarity measures. Second, we proved that the similarity values for the one-norm should be larger than that for the two-norm for her pattern recognition problem. Third, we proposed a more scattered similarity measure to help researchers determine patterns. Our findings may shed light on the ongoing debate between Li and Cheng, 2002, Mitchell, 2003.     

Fabrication of Reduced Fat Products by Controlled Heteroaggregation of Oppositely Charged Lipid Droplets

Abstract: This study describes how the rheological properties of colloidal dispersions formed by heteroaggregation of oppositely charged protein‐coated lipid droplets depend on total particle concentration. Mixed‐particle emulsions were formed by mixing single‐particle emulsions containing either β‐lactoglobulin‐coated lipid droplets (ζ≈?42 mV, d₄₃≈ 0.35 μm) or lactoferrin (LF)‐coated lipid droplets (ζ≈+26 mV, d₄₃≈ 0.32 μm). A series of single‐particle and mixed‐particle emulsions with different total fat contents (5% to 40%) were prepared, and their mean particle size, apparent viscosity, and shear modulus were measured. Mixed‐particle emulsions (40% LF: 60%β‐Lg) containing relatively high fat contents (>10%) had high viscosities and paste‐like properties. These rheological characteristics were attributed to extensive particle aggregation and network formation due to electrostatic attraction between oppositely charged droplets. The viscosities of mixed‐particle emulsions were much higher than those of single‐particle emulsions with equivalent fat contents. Measurements of the color coordinates (L*, a*, b*) of mixed‐particle emulsions with low fat contents showed that they had similar appearances as single‐particle emulsions with high fat contents. This study has important implications for the creation of reduced fat foods with similar sensory qualities as higher fat foods. Practical Application: This study describes how novel food materials can be created by mixing together negatively charged and positively charged lipid droplets. These mixed emulsions may be suitable for the creation of reduced fat products, since they have high viscosities and paste‐like properties at relatively low fat contents.     

Homogeneous cooling of hard ellipsoids

We study the homogeneous cooling of hard, smooth ellipsoids in three dimensions using event driven numerical simulation. The elongation of the particle has a strong effect on the cooling behavior. Weakly elongated ellipsoids display two distinct cooling regimes. For small times, the translational and rotational energy decay at a different rates. Once their ratio reaches a time-independent value (different from equipartition), the overall temperature of the system decays like t ^?2, as predicted by Haff??s law. For more elongated ellipsoids the translational and rotational temperatures rapidly reach a constant ratio near unity. The cooling behavior in the homogeneous state can be predicted from Haff??s law and the equilibrium collision rate.     

Exploring inattention and distraction in the SafetyNet Accident Causation Database

Distraction and inattention are considered to be very important and prevalent factors in the causation of road accidents. There have been many recent research studies which have attempted to understand the circumstances under which a driver becomes distracted or inattentive and how distraction/inattention can be prevented. Both factors are thought to have become more important in recent times partly due to the evolution of in-vehicle information and communication technology. This study describes a methodology that was developed to understand when factors such as distraction and inattention may have been contributors to crashes and also describes some of the consequences of distraction and inattention in terms of subsequent driver actions.     

Nanoemulsion-based delivery systems for poorly water-soluble bioactive compounds: Influence of formulation parameters on Polymethoxyflavone crystallization

Polymethoxyflavones (PMFs) extracted from citrus peel exhibit potent anti-cancer activity, but are highly hydrophobic molecules with poor solubility in both water and oil at ambient and body temperature, which limits their bioavailability. The possibility of encapsulating PMFs within nanoemulsion-based delivery systems to facilitate their application in nutraceutical and pharmaceutical products was investigated. The influence of oil type (corn oil, MCT, orange oil), emulsifier type (β-lactoglobulin, lyso-lecithin, Tween, and DTAB), and neutral cosolvents (glycerol and ethanol) on the formation and stability of PMF-loaded nanoemulsions was examined. Nanoemulsions (r < 100 nm) could be formed using high pressure homogenization for all emulsifier types, except DTAB. Lipid droplet charge could be altered from highly cationic (DTAB), to near neutral (Tween), to highly anionic (β-lactoglobulin, lyso-lecithin) by varying emulsifier type. PMF crystals formed in all nanoemulsions after preparation, which had a tendency to sediment during storage. The size, morphology, and aggregation of PMF crystals depended on preparation method, emulsifier type, oil type, and cosolvent addition. These results have important implications for the development of delivery systems for bioactive components that have poor oil and water solubility at application temperatures.     

Systematic skin segmentation: merging spatial and non-spatial data

Skin detection is used in applications ranging from face detection, tracking of body parts, hand gesture analysis, to retrieval and blocking objectionable content. We present a systematic approach for robust skin segmentation using graph cuts. The skin segmentation process starts by exploiting the local skin information of detected faces. The detected faces are used as foreground seeds for calculating the foreground weights of the graph. If local skin information is not available, we opt for the universal seed. To increase the robustness, the decision tree based classifier is used to augment the universal seed weights when no local information is available in the image. With this setup, we achieve robust skin segmentation, outperforming off-line trained classifiers. The setup also provides a generic skin detection system, using positive training data only. With face detection, we take advantage of the contextual information present in the scene. With the weight augmentation, we provide a setup for merging spatial and non-spatial data. Experiments on two datasets with annotated pixel-level ground truth show that the systematic skin segmentation approach outperforms other approaches and provides robust skin detection.