<p dir="ltr">Land cover classification has always been a crucial topic in the remote sensing domain. Utilizing data collected by unmanned aerial vehicles and satellites, researchers can detect land degradation, monitor environmental changes, and provide insights for urban planning. Recent advancements in large multi-modal models have enabled open-vocabulary classification, which is particularly beneficial in this field. Becuase of the pre-training method, these models can perform zero-shot inference on unseen data, significantly reducing the costs associated with data collection and model training. This open-vocabulary feature of large-scale vision-language pre-training aligns well with the requirements of land cover classification, where benchmark datasets in the remote sensing domain comprise various categories, and transferring results from one dataset to another through supervised learning methods is challenging.</p><p dir="ltr">In this thesis, the author explored the performance of zero-shot CLIP and linear probe CLIP to assess the feasibility of using the CLIP model for land cover classification tasks. Further, the author fine-tuned CLIP by creating hierarchical label sets for the datasets, leading to better zero-shot classification results and improving overall accuracy by 2.5%. Regarding data engineering, the author examined the performance of zero-shot CLIP and linear probe CLIP across different categories and proposed a categorization method for land cover datasets. In summary, this work evaluated CLIP's overall performance on land cover datasets of varying spatial resolutions and proposed a hierarchical classification method to enhance its zero-shot performance. The thesis also offers a practical approach for modifying current dataset categorizations to better align with the model.</p>

This essay examines the communication strategies employed by Jacinda Ardern and Erna Solberg during the Covid 19 crisis, with a focus on how they exhibit feminine leadership at- tribute s that show empathetic and interpersonal skills . W e find that both Ardern and Solberg demonstrated similarities in language and visual elements from their Facebook posts during the selected period 21 03 2020 to18 06 2021 to communicate their main messages. They em- phasize interpersonal communication to establish and maintain relationships by expressing gratitude towards their respective populations. By categorizing the behaviours observed in Ardern and Solberg’s communication, specifically the pract ices of preserving, mutual em- powerment, and team creation as described by Fletcher (1999), skills such as empathy, caring and cooperation were identified . This essay's results indicate that feminine leadership utes, as exemplified by Ardern and Solbe rg, prove effective in crisis management, particu- larly in addressing concerns, fostering empathy, and promoting collective responsibility.

Adolescents use of alcohol and drugs in Sweden today is problematic. The alcohol consumption has decreased amongst adolescents, but it is still an issue. What the younger generation sees in the media has an effect on their view on alcohol and adolescents may assume it is a natural part of life. Drug abuse is not as widespread, but it is still a problem amongst adolescents and the consumption has a negative effect. The purpose of the study is to analyze how the Swedish youth series Young Royals are framing alcohol and drugs amongst adolescents. The study will use a multimodal critical discourse analysis to see what discourses there are about alcohol and drugs. Previous research has discovered that teenage characters in film usually consume alcohol together with others. Consumption of drugs in film usually does not lead to any consequences for the characters. Adolescents are influenced by what media present to the public, especially what they see on film and tv. This study has analyzed the first season of the series Young Royals on Netflix. The purpose of this study was to analyze the scenes containing alcohol or drugs. Scenes where alcohol or drugs were represented, portrayed, mentioned or the characters talked about the subject without expressing it, were analyzed. By applying multimodal critical discourse analysis, this study has been able to see that results from previous studies are similar. The discourses we found in the study are for instance that alcohol is a norm and a natural part of the adolescents social life. The second discourse we found is class-distinction where different classes use alcohol and drugs in different ways. One discourse showed consequences of consuming alcohol and drugs and the most dominant discourse showed positive consequences. Young Royals also had two discourses about drug consumption, one when it was framed as a natural way of partying and one where the consumption was framed as an addiction. At last there were a few discourses of motives for consuming alcohol and drugs in the series. The motives are feeling bad, social, accomplishment and addiction. Some discourses are dominating but Young Royals presents a contrast between positive and negative aspects of alcohol and drugs.

I denna uppsats undersöks om svenska kommuner uppfyller lagarna om tillgänglighetsanpassning på offentliga webbplatser. I dagens digitaliserade samhälle används internet som ett kommunikationsverktyg för att framföra samhällsinformation. På grund av olika varianter av funktionsnedsättningar kan information som publiceras på internet vara svår att ta del av. Genom att en webbplats fullföljer lagarna för tillgänglighetsanpassning på offentliga webbplatser kan information som publiceras nå ut till en bredare målgrupp. Eftersom kommunal information är viktig för alla invånare bosatt i en kommun är det av stor betydelse att invånare har möjlighet att ta del av information som publiceras på kommunala webbplatser. I denna uppsats undersöker vi 10 webbplatser från olika kommuner i olika delar av Sverige; sedan analyseras materialet och framställer om det finns skillnader mellan de olika kommunerna. Undersökningen utförs utifrån europeiska standarden EN 301 549 V3.2.1, baserad på WCAG 2.1. Materialet presenteras med hjälp av diagram samt med beskrivande text. / This essay examines how Swedish municipalities follow the laws on accessibility adaptation on public websites. In today's digitized society, the internet is used as a communication tool to convey social information. Due to different types of disabilities, information published on the Internet can be difficult to read. By complying with the laws for accessibility on public websites, published information can reach a wider target group. As municipal information is important for all residents of a municipality, it is important that residents can get the information that is published on municipal websites. In this essay, we examine 10 websites from different municipalities in different parts of Sweden; then we analyse the material and present whether there are differences between the different municipalities. The examination is carried out based on the European standard EN 301 549 V3.2.1, based on WCAG 2.1. The material is presented using diagrams and descriptive text.

Emotion adds an important element to the discussion of how information is conveyed and processed by humans; indeed, it plays an important role in the contextual understanding of messages. This research is centered on investigating relevant features for affect classification, along with modeling the multimodal and multitemporal nature of emotion. The use of formant-based features for affect classification is explored. Since linear predictive coding (LPC) based formant estimators often encounter problems with modeling speech elements, such as nasalized phonemes and give inconsistent results for bandwidth estimation, a robust formant-tracking algorithm was introduced to better model the formant and spectral properties of speech. The algorithm utilizes Gaussian mixtures to estimate spectral parameters and refines the estimates using maximum a posteriori (MAP) adaptation. When the method was used for features extraction applied to emotion classification, the results indicate that an improved formant-tracking method will also provide improved emotion classification accuracy. Spectral features contain rich information about expressivity and emotion. However, most of the recent work in affective computing has not progressed beyond analyzing the mel-frequency cepstral coefficients (MFCC’s) and their derivatives. A novel method for characterizing spectral peaks was introduced. The method uses a multi-resolution sinusoidal transform coding (MRSTC). Because of MRSTC’s high precision in representing spectral features, including preservation of high frequency content not present in the MFCC’s, additional resolving power was demonstrated. Facial expressions were analyzed using 53 motion capture (MoCap) markers. Statistical and regression measures of these markers were used for emotion classification along the voice features. Since different modalities use different sampling frequencies and analysis window lengths, a novel classifier fusion algorithm was introduced. This algorithm is intended to integrate classifiers trained at various analysis lengths, as well as those obtained from other modalities. Classification accuracy was statistically significantly improved using a multimodal-multitemporal approach with the introduced classifier fusion method. A practical application of the techniques for emotion classification was explored using social dyadic plays between a child and an adult. The Multimodal Dyadic Behavior (MMDB) dataset was used to automatically predict young children’s levels of engagement using linguistic and non-linguistic vocal cues along with visual cues, such as direction of a child’s gaze or a child’s gestures. Although this and similar research is limited by inconsistent subjective boundaries, and differing theoretical definitions of emotion, a significant step toward successful emotion classification has been demonstrated; key to the progress has been via novel voice and visual features and a newly developed multimodal-multitemporal approach.

Den här uppsatsen undersökte hur medielogiken uttrycks i Rädda Barnens och Sverige för UNHCR:s videomaterial under flyktingkrisen september 2015. Då majoriteten av den medialiseringsrelaterade forskningen i Sverige har intresserat sig för partiledardebatter och politisk kommunikation finns det en kunskapslucka kring hur medielogiken formar andra sfärer i samhället. Denna studie avsåg att genom semiotisk analys och mulitmodal analys undersöka videomaterial i dessa två ideella organisationer, för att identifiera mediernas berättartekniker i dem. Sammanlagt analyserades fyra videoklipp i två ideella organisationers Facebook-sidor under september 2015 utifrån teorin om semiotik och medielogik. Två tydliga teman i videoklippen framträdde ur resultatet. Dessa benämnde vi som Flyktingkrisen som empati och Flyktingkrisen som verklighet. Ur dessa framkom fyra underteman. Dessa benämnde vi som dramatisering, dokumentering, reklamformat och nyhetsformat. Studiens resultat visade även att medielogiken personifiering förekommer starkt i samtliga av våra studieobjekt. Utifrån detta resultat kunde vi dra slutsatser kring genom vilka stereotyper personifieringen framkom i.   Vår studie bidrar därmed till forskningsfältet med kunskap om stereotyper av flyktingar. I tidigare forskning av Ylva Brune (1998) urskiljs två sorters stereotyper av flyktingar: offer eller hot. Vidare menar Urban Ericsson (2001) att stereotypen den duktige flyktingen tenderar att förekomma i media endast på individnivå. I vårt material fann vi att en flykting gestaltades annorlunda än de stereotyper vi fann i vår tidigare forskning. Vi valde att kalla den här kategorin flyktingen som målmedveten då denne är fast besluten att ta sig fram trots omständigheterna. Vi fann även en icke tidigare dokumenterad framställning av flyktingar som likartad publiken när det gäller välfärdsbakgrund. / The aim of this paper was to examine how media logic’s are expressed in Save the children’s and Sweden for UNHCR’s video material during the migrant crisis. Since the majority of the mediatization research in the Swedish context has focused on the party leaders and political communication, there is a lack of awareness regarding how media logic shapes other spheres in the community. This study’s objective was to examine video material through semiotic and multimodal analysis, to identify media logics. In total four videos from two nonprofit organizations, that were published in September 2015, were analyzed based on the understanding of media logics.  Two distinctive themes appeared in the results. These were titled migrant crisis as empathy and migrant crisis as reality.  From these, four subthemes appeared. These were titled dramatization, documentation, advertising formats and news format. The result shows that personalization occurs strongly in all of our study.  Our study contributes to the research field with knowledge about stereotypes of refugees. We found a new kind of stereotype that we have chosen to call the determined refugee. The choice of name for this category is based on our interpretation of a refugee as determined to survive. In addition to the study’s contribution of how the representation of refugees among nonprofit organizations now take a new form, we also found a not previously documented representation of refugees like the crowd when it comes to the welfare background.  Our study contributes to knowledge about how mediatization is expressed in strategic communication.

This dissertation argues that music is always ideological. For this thesis two lines of argument are given. The first states that music is always ideological because it requires verbal discourses about itself. The second line of argument states that music is always ideological because it influences the listener affectively. That language is necessary for talk about music is trivial. The point is rather that talk about music is necessary for auditive behaviour to turn into complex cultural artefacts. Without language humans would have no more music than birds, whales or duetting apes. At the other extreme, musical experiences are affective in nature. To have a musical experience is to experience an affective unfolding through time. Affect (as distinguished from the emotions) refers to the amodal properties of perception-such as intensity, shape, rhythm-and lies at the heart of human communication. With its roots in early mother-infant interaction, affective communication is inherently social. Together with discourses about music, the affective properties of musical experiences makes music into an extremely subtle, and thereby efficient, ideological manipulator in various types of social contexts. Finally, the theoretical conclusions reached will be exemplified by introducing a virtual listener, the various facets of whose listening experiences are captured by different analytical methods and listening reports as applied to some of the "serious" music by Frank Zappa. Central for the explanation of these listening experiences are the "passions," that is, the affects, moods and emotions that the music evokes in the listener, or that the listener takes the music to express. / <p>The attached fulltext is a revised version of the original thesis.</p>

KRY is a health-related app that provides a conversation with a doctor or psychologist through an app. KRY has chosen to market themselves through influencer marketing at different platforms like for instance Youtube. When influencers market KRY they encourage their followers to pay money for something that plays an important role regarding their own health. The purpose of this thesis is to answer the questions: What communicational strategies do influencers use to market KRY through sponsored video-vlogs? And how does the brand KRY get communicated through influencer marketing in sponsored video-vlogs? The thesis is considered relevant and is contributing to existing field of research because of the fact that much of the already existing research regarding influencer marketing is examining marketing of physical products. In other words, there exists lots of research regarding influencer marketing, but not much research focusing of influencer marketing of health-related services. This is the knowledge gap that this thesis aims to fill. The method the thesis is using is a multimodal analysis in combination with a rhetorical method. The material consists of the three vlogs in collaboration with KRY on Youtube which have the most views. The result shows that influencers market KRY by focusing on the intimacy between themselves and the viewers, talking about their own experience of using KRY, using emotion-related arguments and customizing the marketing after their own brand and their specific target groups of viewers. KRY as a brand did not get communicated very clearly, but could be distinguished by the fact that the influencers showed the app in their vlogs and that the brand KRY got related to positive emotions and words by all the influencers.

With changes in the outside world and that we are constantly faced with potential risks, the effectiveness of risk communication has become a decisive factor. The purpose of this study is to investigate how the Swedish Civil Protection and Preparedness Authority (MSB) has adopted principles for effective risk communication and how they have designed their communication to best prepare citizens for potential crisis. The study also aims to examine how the communication features an image of the nation and how roles in society during a crisis are represented. To achieve this purpose, a normative and critical multimodal method has been applied to MSB:s risk communication in the form of the brochure "If crisis or war comes". The result of the study has shown that MSB largely adopts the principles of effective risk communication, but some parts of the brochure can be viewed as fairly misleading and have underlying ideas that may lead to undesired effects. The image of the nation in the brochure is produced as rational by describing the authorities and citizens as reasonably and active, which in some ways can be useful in gaining trust from the recipient but can also contradict in getting the people to act.

La biométrie est une technologie émergente qui propose de nouvelles méthodes de contrôle, d’identification et de sécurité. Les systèmes biométriques sont souvent victimes de menaces. La reconnaissance faciale est populaire et plusieurs approches existantes utilisent des images dans le spectre visible. Ces systèmes traditionnels opérant dans le spectre visible souffrent de plusieurs limitations dues aux changements d’éclairage, de poses et d’expressions faciales. La méthodologie présentée dans cette thèse est basée sur de la reconnaissance faciale multispectrale utilisant l'imagerie infrarouge et visible, pour améliorer la performance de la reconnaissance faciale et pallier les insuffisances du spectre visible. Les images multispectrales utilisées cette étude sont obtenues par fusion d’images visibles et infrarouges. Les différentes techniques de reconnaissance sont basées sur l’extraction de caractéristiques telles que la texture et les points d’intérêt par les techniques suivantes : une extraction hybride de caractéristiques, une extraction binaire de caractéristiques, une mesure de similarité tenant compte des caractéristiques extraites. / Biometrics is an emerging technology that proposes new methods of control, identification and security. Biometric systems are often subject to risks. Face recognition is popular and several existing approaches use images in the visible spectrum. These traditional systems operating in the visible spectrum suffer from several limitations due to changes in lighting, poses and facial expressions. The methodology presented in this thesis is based on multispectral facial recognition using infrared and visible imaging, to improve the performance of facial recognition and to overcome the deficiencies of the visible spectrum. The multispectral images used in this study are obtained by fusion of visible and infrared images. The different recognition techniques are based on features extraction such as texture and points of interest by the following techniques: a hybrid feature extraction, a binary feature extraction, a similarity measure taking into account the extracted characteristics.