Generative Part-Based Gabor Object Detector

Object detection is a fundamental task of computer vision that is utilized as a core part in a number of industrial and scientific applications, for example, in robotics, where objects need to be correctly detected and localized prior to being grasped and manipulated. Existing object detectors vary in (i) the amount of supervision they need for training, (ii) the type of a learning method adopted (generative or discriminative) and (iii) the amount of spatial information used in the object model (model-free, using no spatial information in the object model, or model-based, with the explicit spatial model of an object). Although some existing methods report good performance in the detection of certain objects, the results tend to be application specific and no universal method has been found that clearly outperforms all others in all areas. This work proposes a novel generative part-based object detector. The generative learning procedure of the developed method allows learning from positive examples only. The detector is based on finding semantically meaningful parts of the object (i.e. a part detector) that can provide additional information to object location, for example, pose. The object class model, i.e. the appearance of the object parts and their spatial variance, constellation, is explicitly modelled in a fully probabilistic manner. The appearance is based on bio-inspired complex-valued Gabor features that are transformed to part probabilities by an unsupervised Gaussian Mixture Model (GMM). The proposed novel randomized GMM enables learning from only a few training examples. The probabilistic spatial model of the part configurations is constructed with a mixture of 2D Gaussians. The appearance of the parts of the object is learned in an object canonical space that removes geometric variations from the part appearance model. Robustness to pose variations is achieved by object pose quantization, which is more efficient than previously used scale and orientation shifts in the Gabor feature space. Performance of the resulting generative object detector is characterized by high recall with low precision, i.e. the generative detector produces large number of false positive detections. Thus a discriminative classifier is used to prune false positive candidate detections produced by the generative detector improving its precision while keeping high recall. Using only a small number of positive examples, the developed object detector performs comparably to state-of-the-art discriminative methods.

How Lean Six Sigma and Agile principles optimize ITIL-based processes

Imagine the potential implications of an organization whose business and IT processes are well aligned and are capable of reactively and proactively responding to the external and internal changes. The Philips IT Infrastructure and Operations department (I&O) is undergoing a series of transformation activities to help Philips business keeping up with the changes. I&O would serve a critical function in any business sectors; given that the I&O’s strategy switched from “design, build and run” to “specify, acquire and performance manage”, that function is amplified. In 2013, I&O’s biggest transforming programme I&O Futures engaged multiple interdisciplinary departments and programs on decommissioning legacy processes and restructuring new processes with respect to the Information Technology Internet Library (ITIL), helping I&O to achieve a common infrastructure and operating platform (CI&OP). The author joined I&O Futures in the early 2014 and contributed to the CI&OP release 1, during which a designed model Bing Box and its evaluations were conducted through the lens of six sigma’s structured define-measure-analyze-improve-control (DMAIC) improvement approach. This Bing Box model was intended to firstly combine business and IT principles, namely Lean IT, Agile, ITIL best practices, and Aspect-oriented programming (AOP) into a framework. Secondly, the author implemented the modularized optimization cycles according to the defined framework into Philips’ ITIL-based processes and, subsequently, to enhance business process performance as well as to increase efficiency of the optimization cycles. The unique of this thesis is that the Bing Box model not only provided comprehensive optimization approaches and principles for business process performance, but also integrated and standardized optimization modules for the optimization process itself. The research followed a design research guideline that seek to extend the boundaries of human and organizational capabilities by creating new and innovative artifacts. The Chapter 2 firstly reviewed the current research on Lean Six Sigma, Agile, AOP and ITIL, aiming at identifying the broad conceptual bases for this study. In Chapter 3, we included the process of constructing the Bing Box model. The Chapter 4 described the adoption of Bing Box model: two-implementation case validated by stakeholders through observations and interviews. Chapter 5 contained the concluding remarks, the limitation of this research work and the future research areas. Chapter 6 provided the references used in this thesis.

Lignin oxidation by PCD technology

The purpose of this study is to investigate whether commercial Kraft lignin can be treated with pulsed corona discharge apparatus so that it becomes active. Active lignin refers to the kind of lignin that can be precipitated on the surface of a fiber by lowering the pH. A secondary agenda here is to remove the pungent smell of kraft lignin, which is caused by organically bound sulfur. It is expected that the study will identify mild processing conditions and parameters for achievement of the desired outcome. In the literature review, the properties of lignin are explained, as is their impact on any further processing. In addition, a number of processes are described for the oxidation of lignin in a variety of applications. In the experimental part of the study, test runs were conducted to determine the effects of oxygen supply and pulse frequency on oxidation results, where the purpose is to produce reactive lignin and to find a process that is feasible at an industrial scale. Based on the reported experiments, lignin could not be made active or precipitated to the surface of the fiber. Actual changes in the structure of lignin were not observed, but the pungent smell of lignin was removed. The exact reason for this change could not be established because sulfur NMR analysis did not work for the lignin samples.

Stability studies and shelf life estimation of a soy-based dessert

This work is aimed at evaluating the physicochemical, physical, chromatic, microbiological, and sensorial stability of a non-dairy dessert elaborated with soy, guava juice, and oligofructose for 60 days at refrigerated storage as well as to estimate its shelf life time. The titrable acidity, pH, instrumental color, water activity, ascorbic acid, and physical stability were measured. Panelists (n = 50) from the campus community used a hedonic scale to assess the acceptance, purchase intent, creaminess, flavor, taste, acidity, color, and overall appearance of the dessert during 60 days. The data showed that the parameters differed significantly (p < 0.05) from the initial time, and they could be fitted in mathematical equations with coefficient of determination above 71%, aiming to consider them suitable for prediction purposes. Creaminess and acceptance did not differ statistically in the 60-day period; taste, flavor, and acidity kept a suitable hedonic score during storage. Notwithstanding, the sample showed good physical stability against gravity and presented more than 15% of the Brazilian Daily Recommended Value of copper, iron, and ascorbic acid. The product shelf life estimation found was 79 days considering the overall acceptance, acceptance index and purchase intent.

Instrumental color and sensory acceptance of soy-based emulsions: a response surface approach

Response Surface Methodology (RSM) was applied to evaluate the chromatic features and sensory acceptance of emulsions that combine Soy Protein (SP) and red Guava Juice (GJ). The parameters analyzed were: instrumental color based on the coordinates a* (redness), b* (yellowness), L* (lightness), C* (chromaticity), h* (hue angle), visual color, acceptance, and appearance. The analyses of the results showed that GJ was responsible for the high measured values of red color, hue angle, chromaticity, acceptance, and visual color, whereas SP was the variable that increased the yellowness intensity of the assays. The redness (R²adj = 74.86%, p < 0.01) and hue angle (R²adj = 80.96%, p < 0.01) were related to the independent variables by linear models, while the sensory data (color and acceptance) could not be modeled due to a high variability. The models of yellowness, lightness, and chromaticity did not present lack of fit but presented adjusted determination coefficients bellow 70%. Notwithstanding, the linear correlations between sensory and instrumental data were not significant (p > 0.05) and low Pearson coefficients were obtained. The results showed that RSM is a useful tool to develop soy-based emulsions and model some chromatic features of guava-based emulsions through RSM.

Phase transitions in biodegradable films based on blends of gelatin and poly (vinyl alcohol)

The aim of this work was to study the effect of the hydrolysis degree (HD) and the concentration (C PVA) of two types of poly (vinyl alcohol) (PVA) and the effect of the type and the concentration of plasticizers on the phase properties of biodegradable films based on blends of gelatin and PVA, using a response-surface methodology. The films were made by casting and the studied properties were their glass (Tg) and melting (Tm) transition temperatures, which were determined by diferential scanning calorimetry (DSC). For the data obtained on the first scan, the fitting of the linear model was statistically significant and predictive only for the second melting temperature. In this case, the most important effect on the second Tm of the first scan was due to the HD of the PVA. In relation to the second scan, the linear model could be fit to Tg data with only two statistically significant parameters. Both the PVA and plasticizer concentrations had an important effect on Tg. Concerning the second Tm of the second scan, the linear model was fit to data with two statistically significant parameters, namely the HD and the plasticizer concentration. But, the most important effect was provoked by the HD of the PVA.

Chemical pattern of brazilian apples: a chemometric approach based on the Fuji and Gala varieties

The chemical composition of apple juices may be used to discriminate between the varieties for consumption and those for raw material. Fuji and Gala have a chemical pattern that can be used for this classification. Multivariate methods correlate independent continuous chemical descriptors with the categorical apple variety. Three main descriptors of apple juice were selected: malic acid, total reducing sugar and total phenolic compounds. A chemometric approach, employing PCA and SIMCA, was used to classify apple juice samples. PCA was performed with 24 juices from Fuji and Gala, and SIMCA, with 15 juices. The exploratory and predictive models recognized 88% and 64%, respectively, as belonging to a mixed domain. The apple juice from commercial fruits shows a pattern related to cv. Fuji and Gala with boundaries from 0.18 to 0.389 g.100 mL-1 (malic acid), from 8.65 to 15.18 g.100 mL-1 (total reducing sugar) and from 100 to 400 mg.L-1 (total phenolic compounds), but such boundaries were slightly shorter in the remaining set of commercial apple juices, specifically from 0.16 to 0.36 g.100 mL-1, from 9.25 to 15.5 g.100 mL-1 and from 180 to 606 mg.L-1 for acidity, reducing sugar and phenolic compounds, respectively, representing the acid, sweet and bitter tastes.

Mechanical and water vapor permeability properties of biodegradables films based on methylcellulose, glucomannan, pectin and gelatin

Mimic biological structures such as the cell wall of plant tissues may be an alternative to obtain biodegradable films with improved mechanical and water vapor barrier properties. This study aims to evaluate the mechanical properties and water vapor permeability (WVP) of films produced by using the solvent-casting technique from blended methylcellulose, glucomannan, pectin and gelatin. First, films from polysaccharides at pH 4 were produced. The film with the best mechanical performance (tensile strength = 72.63 MPa; elongation = 9.85%) was obtained from methylcellulose-glucomannan-pectin at ratio 1:4:1, respectively. Then, gelatin was added to this polysaccharide blend and the pH was adjusted to 4, 5 and 6. Results showed significant improvement in WVP when films were made at pH 5 and at polysaccharides/gelatin ratio of 90/10 and 10/90, reaching 0.094 and 0.118 g.mm/h.m².kPa as values, respectively. Films with the best mechanical properties were obtained from the blend of polysaccharides, whereas WVP was improved from the blend of polysaccharides and gelatin at pH 5.

Sensory acceptability and physical stability evaluation of a prebiotic soy-based dessert developed with passion fruit juice

The aim of this study was to use a factorial design approach for developing a palatable and stable soy-based dessert with the addition of Soy Protein (SP), oligofructose, and Passion Fruit Juice (PFJ). Panelists (n = 50) used a seven-point hedonic scale to assess the overall liking, degree of liking of creaminess, taste, and color of the desserts. In addition, the samples were submitted to a preference ranking test in order to evaluate the products' preference. Water Holding Capacity (WHC) and backscattering (BS) measures were also determined to assess the physical stability of the trials. Sample F3 (35% PFJ and 2% SP) was the only one that presented a WHC index of 94.8%; moreover, none of the developed samples had synerisis after 72 hours of storage indicating adequate physical stability of the emulsion process. Samples F2 (25% PFJ, and 3.0% SP), F4 (35% PFJ, and 3.0% SP), and F5 (30% PFJ, and 2.5% SP) presented mean hedonic scores above 'slightly liked' for all sensory attributes. The acceptance index of samples varied from 62.50 to 88% showing the great sensory potential of such products.

Application of carnauba-based wax maintains postharvest quality of 'Ortanique' tangor

This study aimed at evaluating compositional changes in the quality of 'Ortanique' tangor after coating with the carnauba-based waxes Aruá Tropical® or Star Light®. The storage conditions studied simulated those of local marketing (22 ± 2 °C, 60 ± 5% RH). Non-destructive analysis, mass loss, peel color, and sensory evaluation, were performed upon coating and every three days up to the fifteenth day of storage. Destructive analysis, peel moisture content, chlorophyll of the peel, pulp color, juice content, soluble solids (SS), titratable acidity (TA), pH, and soluble solids to titratable acidity ratio, were performed upon coating and every four days up to the sixteenth day of storage. The assay was conducted using an entirely randomized design, with three replications (destructive analyses) or ten replications (non-destructive analyses), in a split plot scheme. Wax-coating, especially Aruá Tropical®, maintained fruit freshness by reducing mass loss and peel dehydration and retaining green color. Peel moisture content, chlorophyll content, and juice content had lower rates in the wax coated fruits. Puncture force, soluble solids, titratable acidity, pH, and soluble solids to titratable acidity ratio varied vary little over the course of storage. Sensory evaluation showed that the application of Aruá Tropical keeps 'Ortanique' tangor fresher for 6 days longer for commercialization.

Application of protein-phenolic based coating on tomatoes (Lycopersicum esculentum)

The aim of this study was to investigate the use of protein-phenolic based coating made from fermented rice bran on cherry tomatoes (Lycopersicum esculentum). Tests were performed with glycerol 3% (v/v), glycerol with protein-phenolic rice bran extract (5%), glycerol with protein-phenolic extract after 96 hours of fermentation (5%), and a control (without coating). The coated cherry tomatoes were kept at room temperature for 28 days. Mass loss, pH and acidity, total soluble solids, and carotenoids were determined every 96 hours. The coating made from the biomass extract reduced the carotenoid and acidity levels in the fruits studied by 17 and 21.1%, respectively, compared to the control. The coating proved an efficient barrier to water vapor with mass loss of 57% less than the control suggesting that it can be used as an alternative for vegetable tissue conservation.

Rosemary extract and celery-based products used as natural quality enhancers for colonial type salami with different ripening times

This study aimed to evaluate the use of rosemary (Rosmarinus officinalis) extract (RE), celery (Apium graveolis), and low levels of NO3 and NO2 as natural agents to enhance the quality of colonial salami. Salami was produced according to three treatments: (A) Control: 0.1% curing salt; (B) Rosemary: 0.05% curing salt + 0.5% RE (rosemary extract); and (C) Rosemary+celery: 0.14% Veg 503 + 0.27% Veg 504 (sea salt plus celery) + 0.5% of RE (rosemary extract). There was no effect (P > 0.05) of the treatments on water activity, Na content, and residual NO3 and NO2. Fatty acids C18:2 and C20:4 were reduced (P < 0.05) during the ripening period in the Control treatment indicating possible oxidation. The use of celery resulted in lower pH values (P < 0.05) in the salami. Reduced addition of NO3 and NO2 resulted in salami lighter in color (higher L* values, P < 0.05) at the 12th day of ripening. In conclusion, celery-based products proved to be an effective source of NO2 and NO3 for color development, but the low pH of the product indicates the need for better evaluation of its use in fermented salami. The RE (rosemary extract) reduced fat oxidation in salami, but this needs further evaluation.

Stability of cassava flour-based food bars

The consumption of Brazilian cassava has been reduced due to a lack of adjustment to the modern lifestyle. To reverse this trend, new products could be developed specifically targeted to high-value niche markets. Cereal bars stand out as fast food high in nutritional value. A bar formula mimicking cereal bars was prepared using a mixture of Brazilian cassava flour, hydrogenated vegetable fat, dried bananas, ground cashew nuts, and glucose syrup. After being pressed, the bars were dried for 1 hour at 65 °C, packaged in films, and stored under ambient conditions. Its stability was continuously monitored for 210 days in order to ensure its safety and enable its introduction to the market. Texture loss was observed in the packed bars after 90 days of storage, but the sensory characteristics allowed the testers to perceive this tendency after only 30 days of storage. However, chemical, physical, and microbial analyses confirmed that the bars were safe for consumption for 180 days. The results showed that a 45 g cassava flour-based bar enriched with nuts and dried fruits can meet 6% of the recommended daily fiber intake with a caloric value between that of the common cereal bar and that of an energy bar. Adapting the formula with ingredients (fruits, nuts) from different regions of Brazil may add value to this traditional product as a fast food.

Characterization and classification of pequi trees (Caryocar brasiliense Camb.) based on the profile of volatile constituents using headspace solid-phase microextraction - gas chromatography - mass spectrometry and multivariate analysis

In order to determine the variability of pequi tree (Caryocar brasiliense Camb.) populations, volatile compounds from fruits of eighteen trees representing five populations were extracted by headspace solid-phase microextraction and analyzed by gas chromatography-mass spectrometry. Seventy-seven compounds were identified, including esters, hydrocarbons, terpenoids, ketones, lactones, and alcohols. Several compounds had not been previously reported in the pequi fruit. The amount of total volatile compounds and the individual compound contents varied between plants. The volatile profile enabled the differentiation of all of the eighteen plants, indicating that there is a characteristic profile in terms of their origin. The use of Principal Component Analysis and Cluster Analysis enabled the establishment of markers (dendrolasin, ethyl octanoate, ethyl 2-octenoate and β-cis-ocimene) that discriminated among the pequi trees. According to the Cluster Analysis, the plants were classified into three main clusters, and four other plants showed a tendency to isolation. The results from multivariate analysis did not always group plants from the same population together, indicating that there is greater variability within the populations than between pequi tree populations.

Development and chemical and sensory characterization of pumpkin seed flour-based cereal bars

Pumpkin (Cucurbita maxima), popularly known as squash, is a widely grown vegetable in Brazil. In this study, pumpkin seed flours (PSF) with different granulometries were used: PSF 1 (medium granulometry) and PSF 2 (coarse granulometry) in the preparation of cereal bars (CB) with different combinations with brown oats. Five formulations were prepared: CB-1 (control - 25% brown oats and 0% PSF); CB-2 (12.5% PSF 1 and 12.5% brown oats); CB-3 (25% PSF 1 and 0% brown oats); CB-4 (12.5% PSF 2 and 12.5% brown oats); and CB-5 (25% PSF 2 and 0% brown oats). The acceptance test results were analyzed in a conventional preference mapping which indicated that the bars CB-2 and CB-5 received mostly the maximum hedonistic score. With the objective of developing a cereal bar replacing oats with PSF, the bars CB-2 and CB-5 were compared to the conventional bar CB-1. The cereal bars CB-2 and CB-5 showed an increase in crude protein (87.5% and 62.5%) and in dietary fiber (77% and 44%), respectively. These results allowed the classification of CB-2 and CB-5 as fiber sources; they can, therefore, be classified as light products according to the Brazilian legislation.