This paper deals with the manufacture and characterisation of cheese analogues, which are products, where the milk components are totally or partially replaced by non-milk components. Model samples, edam type cheese and the analogue containing hazelnut oil, were manufactured using the standard way at Brno Mendelu. The main parameters chosen to describe the quality of samples were free and bound fatty acids, volatile compounds and sensory quality. The main aim was to judge the influence of added hazelnut oil on mentioned parameters. For the identification and semiquantification of volatile compounds HS-SPME-GC-MS method was used. Fat extraction from sample was made with a mixture of two solvents (diethylether, petrolether), fatty acids were identified and quantified by GC-FID, with their former acid transesterification using methanolic solution of BF3 as a catalyst. For sensory analysis of the appearance, colour, texture, aroma, taste and overall acceptance of the sample descriptive scales and comparison with standard were used, the structure was observed by electron microscopy. The results show that addition of hazelnut oil affects the sensory properties, ripening process (higher amount of volatile compounds, faster lipolysis), nutrition value (higher amount of unsaturated fatty acids), and the stability of the product.

The presented diploma thesis deals with the extraction and storage stability of lipophilic pigments produced by selected strains of yeasts and algae. In this thesis, there was studied the influence of the selected solvents on the efficiency of extraction, as well as the effect of ambient temperature on the stability of the pigments during storage. The work is divided into two main parts, theoretical and practical part. In the theoretical part knowledge about algae, yeasts and their cultivation parameters is summarized. Furthermore, some information regarding the selected lipid metabolites, their properties and possibilities of application in various branches of industry is mentioned. The practical part deals with the preparation of extracts and stability tests. Extracts were prepared from selected biomass samples in solvents suitable for applications in food industry or cosmetics (ethanol and hexane). Subsequently, the long-term stability tests lasting 4 months and short-term stability tests lasting a total of 28 days were performed on these extracts. The pigments were determined by HPLC and spectrophotometrically, the fatty acid content was determined by GC. For the extraction of pigments from biomass, in the most cases ethanol appears to be the optimal solvent. However, for lipid extraction from biomass, hexane appears to be the optimal solvent for a significant number of samples. In most samples, storage in the freezer showed the most favourable effect on pigment stability, but some samples showed comparable stability even when stored in the refrigerator.

The presented thesis deals with the monitoring sensory quality of processed cheese analogues, focusing mainly on the taste (flavour) and related content of volatile (aroma active) substances. Model samples of analogues were produced by a standard procedure at Tomas Bata University in Zlín. The experimental part was divided into two experiments, which differed in the composition of model samples of analogues. In the first experiment, the traditional fat (butter) was completely replaced by selected vegetable fats (palm, coconut, mixed), in the second experiment only a part (1% w/w - expressed on the total weight of the sample) of butter was replaced by vegetable oils (apricot, flax seed, currant, grape seed). Solid phase microextraction in conjunction with gas chromatography with flame ionization detection was used to determine volatiles. Methods based on valid international standards were used to evaluate the sensory quality of samples, focusing mainly on flavor, namely evaluation using scales (ISO 4121), profile test (EN ISO 13299) and ranking test (ISO 8587). The aim of the work was to assess the effect of the addition of various vegetable fats/oils on the above parameters, at the same time their changes were monitored during 6 months of storage (at 6 ° C). The results showed that the vegetable oil used affects both the sensory quality and the content and composition of volatile substances of analogues. Significant differences between samples were mainly in taste, aroma and overall acceptability. The decreasing overall acceptability of the samples can be expressed in experiment I by the order: product with butter coconut palm mixed fat; in the case of experiment II: product with butter apricot = flax = grape currant oil. Mixed fat, currant and grape seed oil proved to be unsuitable for the production of analogues, on the contrary, analogues with coconut fat and apricot oil were evaluated best. Based on the results of sensory analyses, it was shown that the samples maintain a good sensory quality min. for 3 months while maintaining a low storage temperature ( 6 ° C). An analogue with coconut fat, whose taste, aroma and acceptability were rated as very good, and apricot oil, whose taste, aroma and acceptability were even rated as excellent, could enrich the food offer on the market.

The aim of this master‘s thesis was to determine the aroma active compounds and monitor the sensory quality of apples. The theoretical part deals with the description, composition and properties of apples, information about the aroma compounds, their determination and possibilities for sensory evaluation of apples is also given. The gas chromatography method combined with the solid-phase microextraction, utilized for the specification of aroma compounds, is described. In the experimental part, the SPME-GC-FID method was applied on selected cultivars of apples specially cultivated against apple scab (Venturia inaequalis), i.e. 24 genotypes and Golden Delicious as the control variety. In total 60 various aroma compounds were identified, 22 alcohols, 12 aldehydes, 11 ketones, 10 esters and 5 organic acids. The number and concentration of compounds identified differed according to the cultivars. Sensory analysis was performed using category ordinal scales and the profile test.

The field of image synthesis has seen significant progress recently, including great strides with generative models like Generative Adversarial Networks (GANs), Diffusion Models, and Transformers. These models have shown they can create high-quality images from a variety of text prompts. However, a comprehensive analysis that examines both their performance and possible biases is often missing from existing research. In this thesis, I undertake a thorough examination of several leading text-to-image models, namely Stable Diffusion, DALL-E Mini, Lafite, and Ernie-ViLG. I assess their performance in generating accurate images of human faces, groups, and specified numbers of objects, using both Frechet Inception Distance (FID) scores and R-precision as my evaluation metrics. Moreover, I uncover inherent gender or social biases these models may possess. My research reveals a noticeable bias in these models, which show a tendency towards generating images of white males, thus under-representing minorities in their output of human faces. This finding contributes to the broader dialogue on ethics in AI and sets the stage for further research aimed at developing more equitable AI systems. Furthermore, based on the metrics I used for evaluation, the Stable Diffusion model outperforms the others in generating images from text prompts. This information could be particularly useful for researchers and practitioners trying to choose the most effective model for their future projects. To facilitate further research in this field, I have made my findings, the related data, and the source code publicly available. / Master of Science / In my research, I explored how cutting-edge computer models, namely Stable Diffusion, DALL-E Mini, Lafite, and Ernie-ViLG, can create images from text descriptions, a process that holds exciting possibilities for the future. However, these technologies aren't without their challenges. An important finding from my study is that these models exhibit bias, e.g., they often generate images of white males more than they do of other races and genders. This suggests they're not representing our diverse society fairly. Among these models, Stable Diffusion outperforms the others at creating images from text prompts, which is valuable information for anyone choosing a model for their projects. To help others learn from my work and build upon it, I've made all my data, findings, and the code I used in this study publicly available. By sharing this work, I hope to contribute to improving this technology, making it even better and fairer for everyone in the future.

Also aged bottles are safer to use than new bottles because their chemical leaching was found to be lower than that of new bottles. This study recommends the reassessment of the absence of international guidelines for acetaldehyde in water and foods. The study also recommends that the amount of acetaldehyde that can be added to soft drinks as flavouring agent should be below the specific migration limit (SML) for migration of acetaldehyde from PET bottle into bottle contents. This is essential since the SML was designed to ensure that exposure to acetaldehyde, as a result of intake of bottled water and soft drinks in PET bottles, is below the tolerable daily intake (TDI) for acetaldehyde. As antimony was reported to go beyond the safe limits in some Nigerian bottled water and soft drinks after 11 months of storage this study discourages the use of bottle contents stored for a very long time.

Background: Generative Adversarial Networks (Goodfellow et al., 2014) (GANs)are the current state of the art machine learning data generating systems. Designed with two neural networks in the initial architecture proposal, generator and discriminator. These neural networks compete in a zero-sum game technique, to generate data having realistic properties inseparable to that of original datasets. GANs have interesting applications in various domains like Image synthesis, 3D object generation in gaming industry, fake music generation(Dong et al.), text to image synthesis and many more. Despite having a widespread application domains, GANs are popular for image data synthesis. Various architectures have been developed for image synthesis evolving from fuzzy images of digits to photorealistic images. Objectives: In this research work, we study various literature on different GAN architectures. To understand significant works done essentially to improve the GAN architectures. The primary objective of this research work is synthesis of plant images using Style GAN (Karras, Laine and Aila, 2018) variant of GAN using style transfer. The research also focuses on identifying various machine learning performance evaluation metrics that can be used to measure Style GAN model for the generated image datasets. Methods: A mixed method approach is used in this research. We review various literature work on GANs and elaborate in detail how each GAN networks are designed and how they evolved over the base architecture. We then study the style GAN (Karras, Laine and Aila, 2018a) design details. We then study related literature works on GAN model performance evaluation and measure the quality of generated image datasets. We conduct an experiment to implement the Style based GAN on leaf dataset(Kumar et al., 2012) to generate leaf images that are similar to the ground truth. We describe in detail various steps in the experiment like data collection, preprocessing, training and configuration. Also, we evaluate the performance of Style GAN training model on the leaf dataset. Results: We present the results of literature review and the conducted experiment to address the research questions. We review and elaborate various GAN architecture and their key contributions. We also review numerous qualitative and quantitative evaluation metrics to measure the performance of a GAN architecture. We then present the generated synthetic data samples from the Style based GAN learning model at various training GPU hours and the latest synthetic data sample after training for around ~8 GPU days on leafsnap dataset (Kumar et al., 2012). The results we present have a decent quality to expand the dataset for most of the tested samples. We then visualize the model performance by tensorboard graphs and an overall computational graph for the learning model. We calculate the Fréchet Inception Distance score for our leaf Style GAN and is observed to be 26.4268 (the lower the better). Conclusion: We conclude the research work with an overall review of sections in the paper. The generated fake samples are much similar to the input ground truth and appear to be convincingly realistic for a human visual judgement. However, the calculated FID score to measure the performance of the leaf StyleGAN accumulates a large value compared to that of Style GANs original celebrity HD faces image data set. We attempted to analyze the reasons for this large score.

There are 209 PCB congeners, 136 of which have been found in technical PCB mixtures and hence may be found in the environment as a result of either intentional or unintentional release. The identification and quantification of the congeners are difficult due to analytical bias from coeluting PCBs and other persistent organic pollutants. Among the 209 possible PCB congeners, 19 tri- and tetra-ortho chlorinated congeners exist in stable atropisomeric conformations. The racemization barrier were determined for twelve of the nineteen atropisomers and was found to be between 176-185 kJ × mol-1 and ca. 250 kJ × mol-1 for tri- and tetra-ortho PCB, respectively. Further, a buttressing effect of 6.4 kJ × mol-1 was observed for congeners with vicinal ortho-meta chlorines. Comprehensive two-dimensional gas chromatography (GC×GC) was used to analyze the atropisomers and other PCBs. A Longitudinally Modulated Cryogenic System (LMCS) was used with liquid CO2 as cryogen. The LMCS was optimized for semi-volatile organic substances, primarily PCBs. The trap temperature was shown to be an important factor for the trapping and desorption efficiency, as was the thermal mass of the column used in the modulator region. A number of column sets were tested and the separation efficiency, congener resolution and analysis time was evaluated. Good separation of non- and mono-ortho PCBs and “bulk” PCBs (in a technical PCB) was obtained within 8 min using a smectic liquid crystal column (LC50) as the first and a nonpolar column as the second dimension column. Using a second column, an efficient nonpolar (DB-XLB) column, which separates many PCB congeners, were combined with a polar (cyanopropyl) or shape selective (LC50) second dimension column. As a maximum, 181 of the 209 congeners and 126 of the 136 Aroclor PCBs were resolved. The seven frequently measured PCBs (PCBs 28, 52, 101, 118, 138, 153 and 180) and all WHO-PCBs were separated from all other Aroclor PCBs. Chiral PCBs are released into the environment as racemic mixtures. However, organisms have been shown to enantiomerically enrich many of the atropisomers, suggesting that enantioselective biotransformations occur. Non-racemic PCB enrichment has also been seen in mammalians including humans, which is of particular concern because of the potential health risk. An analytical procedure were therefore developed and used to determine the levels of atropisomeric PCBs, planar-PCBs (WHO-PCBs) and total PCBs in seals with different health status. GC×GC was used to separate the target PCBs from other PCBs and potential interferences. A chiral column (permethylated â-cyclodextrin) was used in combination with a polar or shape selective column and enantiomeric fractions (EFs) were determined for five atropisomeric PCBs, i.e. CBs 91, 95, 132, 149 and 174. Some atropisomers had EF that deviated largely from racemic. The deviation was larger in liver than blubber, indicating enantioselective metabolism. However, there was no selective passage of the studied atropisomeric PCBs across placenta and no selective blood-brain barrier. Similarly, no correlation between EFs and health status was observed, although there was a correlation between total PCBs and health status.

The dairy food industry is continuously striving towards products with higher quality and longer shelf-life available to the customer at low prices. Arla Foods in Linköping, Sweden, is currently investigating the possibilities of changing the material in yogurt packaging containers by replacing the currently used carton with a different and cheaper alternative. A successful switch will give the company an economical advantage without affecting the sensory attributes (smell, taste, sight, and consistency), aroma profile or other important trademarks of the yogurt. This study is designed to examine and compare yogurt that has been stored in different packaging cartons, one coated with a single-layered low-density polyethylene (LDPE) and one coated with a currently used multi-layered ethylene-vinyl alcohol (EVOH). The study was based on the analysis and measurement of sensory attributes performed by experts, physical properties in laboratory and chemical composition in GC-FID/MS together with a discriminative test where a group of people would identify any difference between the yogurts. Together, these analyses would provide an explanation about any differences between the packaging materials by connecting physical, chemical and/or sensory characteristics. The collected results would give a better and more comprehensive picture than each analysis would do separately. The results from the study show that there is a difference between yogurts stored in LDPE-based containers and yogurts stored in EVOH-based containers and that the product was chemically affected, mainly by the level of oxygen in contact with the food. The overall assessment is that the largest difference was discovered in the taste.

This paper reports the study of composition and variation of the chemical components of essential oils of Croton species of Sergipe emphasizing its importance chemotaxonomy. Essential oils from fresh aerial parts obtnained by hydrodistillation from samples of Croton blanchetianus (CB-GA), C. campestris (CC-AML), C. heliotropiifolius collected at different locations in the Sergipe (CH-SG1, CH-SG2, CH-GA, and CH-AML), C. pedicellatus (CP-SI), C. tetradenius (CT-SG) and seven other unidentified taxa, called CSP1-CPS7, were analyzed by GC-MS and GC-FID, such analysis revealed a highly variable chemical composition. It were identified 131 components by comparing the retention indices and mass spectra of compounds with literature data and NIST and Willey mass spectra libraries. With the chemical analysis of oil of Croton were established correlations between the taxa of C. heliotropiifolius, which can be differentiated by using their main chemical markers: (E)-caryophyllene and bicyclogermacrene. For comparative analysis of the chemical composition of essential oils of Croton species, chemometric tools as Principal Component Analysis (PCA) and hierarchic Cluster Analysis (HCA) was applied, by which it was possible to differentiate the components of the essential oil of aerial parts of Croton species grouped by class of compounds according to their biosynthetic origin (monoterpene hydrocarbons, oxygenated monoterpenes, sesquiterpene hydrocarbons, oxygenated sesquiterpenes, phenylpropanoids and derivatives of fatty acids). With this analysis it was possible to form three groups (I, II and III), corroborating the comparative analysis of chromatographic profiles. To our knowledge, this is the first report on the chemical composition of essential oils of C. pedicellatus and C. tetradenius. / Este trabalho relata o estudo da composição e variação dos componentes químicos do óleo essencial das espécies de Croton do estado de Sergipe dando ênfase a sua importância quimiotaxonômica. Os óleos essenciais das partes aéreas frescas obtidos por hidrodestilação das amostras de Croton blanchetianus (CB-GA), C. campestris (CC-AML), C. heliotropiifolius coletadas em diferentes localidades do Estado (CH-SG1, CH-SG2, CH-GA, CH-SI e CH-AML), C. pedicellatus (CP-SI), C. tetradenius (CT-SG) e outros sete táxons ainda não identificados denominados (CSP1-CPS7) foram analisados por CG-EM e CG-DIC revelando uma composição química bastante variável. Foram identificados 131 componentes, baseados nos índices de retenção e na comparação de seus espectros de massas com os da literatura e com as bibliotecas de espectros de massas Willey e NIST. Com a análise química dos óleos de Croton foram estabelecidas correlações entre os táxons de C. heliotropiifolius, os quais foram diferenciados dos demais mediante a utilização de seus marcadores químicos principais: (E)-cariofileno e biciclogermacreno. Para as análises comparativas da composição química dos óleos essenciais das espécies de Croton aplicaram-se ferramentas quimiométricas de Análise dos Componentes Principais (ACP) e Análise de Agrupamentos Herarquicos (AAH), pelas quais foi possível diferenciar os componentes do óleo essencial das partes aéreas das espécies de Croton agrupados por classe de compostos de acordo com sua origem biossintética (hidrocarbonetos monoterpênicos, monoterpenos oxigenados, hidrocarbonetos sesquiterpênicos, sesquiterpenos oxigenados, fenilpropanóides e derivados de ácidos graxos). Com essas análises foi possível a formação de três grupos (I, II e III), corroborando com as análises comparativas dos perfis cromatográficos. Ao nosso conhecimento, este é o primeiro relato sobre a composição química dos óleos essenciais de C. pedicellatus e C. tetradenius.