Content-based video retrieval research combines expertise from many different areas, such as signal processing, machine learning, pattern recognition, and computer vision. As video extends into both the spatial and the temporal domain, we require techniques for the temporal decomposition of footage so that specific content can be accessed. This content may then be semantically classified - ideally in an automated process - to enable filtering, browsing, and searching. An important aspect that must be considered is that pictorial representation of information may be interpreted differently by individual users because it is less specific than its textual representation. In this thesis, we address several fundamental issues of content-based video retrieval for effective handling of digital footage. Temporal segmentation, the common first step in handling digital video, is the decomposition of video streams into smaller, semantically coherent entities. This is usually performed by detecting the transitions that separate single camera takes. While abrupt transitions - cuts - can be detected relatively well with existing techniques, effective detection of gradual transitions remains difficult. We present our approach to temporal video segmentation, proposing a novel algorithm that evaluates sets of frames using a relatively simple histogram feature. Our technique has been shown to range among the best existing shot segmentation algorithms in large-scale evaluations. The next step is semantic classification of each video segment to generate an index for content-based retrieval in video databases. Machine learning techniques can be applied effectively to classify video content. However, these techniques require manually classified examples for training before automatic classification of unseen content can be carried out. Manually classifying training examples is not trivial because of the implied ambiguity of visual content. We propose an unsupervised learning approach based on latent class modelling in which we obtain multiple judgements per video shot and model the users' response behaviour over a large collection of shots. This technique yields a more generic classification of the visual content. Moreover, it enables the quality assessment of the classification, and maximises the number of training examples by resolving disagreement. We apply this approach to data from a large-scale, collaborative annotation effort and present ways to improve the effectiveness for manual annotation of visual content by better design and specification of the process. Automatic speech recognition techniques along with semantic classification of video content can be used to implement video search using textual queries. This requires the application of text search techniques to video and the combination of different information sources. We explore several text-based query expansion techniques for speech-based video retrieval, and propose a fusion method to improve overall effectiveness. To combine both text and visual search approaches, we explore a fusion technique that combines spoken information and visual information using semantic keywords automatically assigned to the footage based on the visual content. The techniques that we propose help to facilitate effective content-based video retrieval and highlight the importance of considering different user interpretations of visual content. This allows better understanding of video content and a more holistic approach to multimedia retrieval in the future.

In an era where Instagram is the new dominating social media platform to reach and communicate with consumers, the demands on companies to differentiate their social media content have increased. Executives seem to avoid social media due to absence of how to manage and learn from it. Therefore, additional in-depth knowledge of how to place apparel products in favourable contexts through images could make marketing efforts more efficient on Instagram. The path this dissertation follows is qualitative with an abductive approach, since the aim of this dissertation is to create an in-depth understanding of what image attributes influence and motivate consumers on Instagram. Primary data have been produced through three semi-structured focus group interviews along with secondary data collected from five apparel brands’ Instagram accounts. This thesis takes its ground in consumer behaviour and theories of visual content, but does not rely on theory alone, as it would have prevented findings of new insights. Findings of previous research on how visual content influence consumers reinforced our research study even when displayed in the forum of Instagram. Additionally, the results of this dissertation stress the importance of placing the product in a relevant context, to a visually appealing background and preferably shown on a human being. These findings may be useful as guidance for apparel companies using Instagram as a promotional tool, as well as for companies who are planning to do so.

Este trabajo de investigación tiene como objetivo demostrar que la importancia la relación de las estrategias de marketing digital con respecto a la intención de compra de planes nutricionales en Lima metropolitana. Nos enfocaremos en tres marcas en específico, ya que se encuentren de manera digital. Además, el contenido visual que manejan y la información nutricional que ofrecen a sus clientes para la captación y venta de sus productos. A lo largo de la investigación se darán detalles de diversas opiniones del público objetivo primarios, así como expertos de la categoría. De manera que se identificaran los principales hallazgos para que las marcas de planes nutricionales puedan reconocer y trabajar en ello en una mejor estrategia digital. / This research work aims to demonstrate the importance of the relationship of digital marketing strategies with respect to the purchase intention of nutritional aircraft in metropolitan Lima. We will focus on three specific brands, as they are found digitally. In addition, the visual content it handles and the nutritional information it offers to its customers for the acquisition and sale of its products. Throughout the investigation details will be given of various opinions of the primary target audience, as well as experts from the category. In order to identify the main findings for brands of nutritional plans, they will be able to recognize and work on a better digital strategy. / Trabajo de investigación

This paper explores the visual content of the Russian Embassy in London in connection to the Skripal poisoning in Salisbury, UK on March 4, 2018. Considering the Skripal poisoning as a significant event in the relationship between the UK and Russia, that Russian disinformation increases substantially surrounding events perceived as key to security, and that the Russian Embassy in London is increasingly coming under investigation for controversial content, the aim of this paper is to deconstruct the tweets made by the Russian Embassy in London using the Skripal poisoning as a case study. This paper also situates this event within the broader context of Russian strategic narratives and disinformation. Tweets made by the Russian Embassy in London were collected between March 4th and Sept. 28th , 2018 and categorized visually into representations of actors, events, and message. Bleiker's (2015) visual content analysis framework was used, specifically through compositional, semiotic and discursive analyses. Through this deconstruction of visual conflict framing, it is hoped that a better understanding of the building blocks of Russian state disinformation can be ascertained within the UK domestic media environment.

Machine-driven analysis of visual data is the hard core of intelligent surveillance systems. Its main goal is to recognize di erent objects in the video sequence and their behaviour. Such operation is very challenging due to the dynamic nature of the scene and the lack of semantic-comprehension for visual data in machines. The general ow of the recognition process starts with the object extraction task. For so long, this task has been performed using image segmentation. However, recent years have seen the emergence of another contender, image matting. As a well-known process, matting has a very rich literature, most of which is designated to interactive approaches for applications like movie editing. Thus, it was conventionally not considered for visual data analysis operations. Following the new shift toward matting as a means to object extraction, two methods have stood out for their foreground-extraction accuracy and, more importantly, their automation potential. These methods are Closed-Form Matting (CFM) and Spectral Matting (SM). They pose the matting process as either a constrained optimization problem or a segmentation-like component selection process. This di erence of formulation stems from an interesting di erence of perspective on the matting process, opening the door for more automation possibilities. Consequently, both of these methods have been the subject of some automation attempts that produced some intriguing results. For their importance and potential, this thesis will provide detailed discussion and analysis on two of the most successful techniques proposed to automate the CFM and SM methods. In the beginning, focus will be on introducing the theoretical grounds of both matting methods as well as the automatic techniques. Then, it will be shifted toward a full analysis and assessment of the performance and implementation of these automation attempts. To conclude the thesis, a brief discussion on possible improvements will be presented, within which a hybrid technique is proposed to combine the best features of the reviewed two techniques. / Thesis / Master of Applied Science (MASc)

För att ge insikt hur design för ett nytt rockalbum ska ske och ge information kring vad som bör finnas med på ett nytt omslag har denna studie utförts. Studien framförde information om att albumsgrafik var relevant. Först sades det att en nedåtgående trend gick att anta, eftersom den digitala musikbranschen tog över från den fysiska. Men det visade sig därefter vara precis lika stor försäljning på de fysiska albumen som de digitala. För att besvara målen i studien utfördes en visuell innehållsanalys och informationssökning på internet. Studiens syfte var att visuellt granska rockalbum från tre olika ursprungsländer, som var Sverige, Storbritannien och USA. Genom analysen kategoriserades designelement, teman och framtogs landstypiska albumomslag. Studien berörde även vem som styr designen av ett omslag samt ifall det kunde ses någon tongivande designer bland de album som ingick i analysen. I slutsatsen framgick det att Storbritannien och USA hade mycket lika grafiskt manér och att omslag från Sverige såg annorlunda ut. Det framgick även att banden styr utseendet på omslaget mer än designern, och från analysen framgick inte någon tongivande designer. / This study is supposed to shed light on the design process behind album covers in the genre rock and give information about what should be displayed on them. Herein it is presented information about how the graphics on a record cover was relevant and didn’t follow the otherwise assumed declining trend. To be able to answer the goals set for the study, a visual content analysis was done and complementary information was gathered via searches on the internet. The aim with the study was to visually analyze cover art in the genre rock from three different countries of origin, these were Sweden, Great Britain and USA. Via the visual content analysis, categories and themes were structured. There were also created three different album covers, one for each country of origin that were based on the information of what should typically exist on the cover. Herein it is shown that designs from Great Britain and USA are much alike, whereas Sweden has a different style. In the study there was presented who decides the design of a new album and also concludes that from the albums in this content analysis there cannot be proven any designer who sets the generic tone for others.

Böcker är, trots ett generellt minskat intresse, fortfarande en viktig del av vårtkulturella samhälle. Att böckers utseende anpassas efter sin målgrupp är ingetovanligt men frågan är om även könet på författaren spelar in gällandeutformningen av omslaget.Syftet med studien var därför att genom en visuell innehållsanalys av böckerspocketomslag kartlägga de skillnader som återfinns på manliga respektivekvinnliga författares omslag.Denna undersökning har baserats på en visuell innehållsanalys där 200pocketomslag analyserades, 100 från svenska författare och 100 från engelskaförfattare. Dessa var därefter jämlikt fördelade mellan könen. I analysenreflekteras det över pocketböckernas motiv, dominerande färg, typsnittsfamiljersamt typsnittsfärger.Resultatet från undersökningen tyder på att kvinnliga författare har ettfemininare utseende på sina böcker överlag, och att svenska kvinnligaförfattares omslag uppvisade flest tecken på en femininare utformning. Det gårdock inte att utesluta att böckernas genre kan ha ett inflytande gällandeböckernas utformning, då det var svårt att precisera genren på många böckeröverlag. / Books are, despite a decline in interest, still an important part of our culturalsociety. It is not unusual for a book’s cover to adapt to its target audience butthe question stands whether the gender of the author functions as a part of thedesign equation as well.The purpose of this study was to through a visual analysis of paperback bookcovers map out differences found between male and female writers’ bookcovers.This study is based on a visual content analysis were 200 paperback coverswere analysed, 100 from Swedish authors and 100 from English authors. Thesewere then evenly divided between the genders. The analysis reflects over themotifs on the covers as well as the dominant colour, font families and fontcolours.The results from the analysis indicate that female authors generally got a morefeminine appearance on their covers, and that covers by Swedish female authorsshowed most signs of a feminine design. Alas, it is impossible to rule out genreas an important influence regarding the books’ cover design. However, as it washard to pinpoint the genre of many of the books analysed it was difficult tomake a distinct connection between genres and books’ cover design.

The constantly evolving use of social media as a marketing tool is commonly known, and the opportunities for companies to get feedback about how successful their marketing is, have grown just as much. Today's companies take great financial risks when choosing marketing material in the traditional marketing way, but as the social network Instagram is free to use, the biggest risk here is to lose followers, and thus potential customers. This study aims to investigate what consumers actually want to see in the direct marketing they themselves have chosen to receive; what the consumers find most appealing and what makes them feel the greatest desire to buy. For this study seven semi-structured interviews were held with seven women all in the age between 16 and 25. They shared their everyday habits on the social media platform Instagram and were shown 15 Instagram pictures from five major fashion companies. The women expressed their spontaneous feelings when the pictures were shown, just as if the images had been found in their own Instagram feed. The images shown consisted of two, by us, defined messages divided into a five-degree scale, where the scale was ranked accordingly: extremely rational, rational, mixed, emotional and extremely emotional messages. The interviewees also commented on their preferences regarding the content of the pictures, if it was up to their own choosing. After analyzing all interviews, we have come to the conclusion that both the rational and the emotional image categories are preferable to the others. These two stand out from the five-degree scale as both more appealing and creating a greater desire to buy. The extreme ends of each category does not suit marketing on Instagram for different reasons. With this information, we hope to help companies with their marketing strategy on Instagram. This thesis is written in Swedish. / Användningen av sociala medier som marknadsföringsverktyg växer ständigt, och så även möjligheterna för företag att få återkoppling angående hur framgångsrik deras marknadsföring är. Dagens företag tar stora ekonomiska risker när de väljer marknadsföringsmaterial på det traditionella sättet, men när det på sociala mediet Instagram är gratis att skapa ett konto blir den största risken att förlora följare och därmed potentiella kunder. Den här studien undersöker vad konsumenter faktiskt önskar att se i den direktreklam de självmant väljer att följa, vad som mest tilltalar dem och vad som får dem att känna mest köplust. Studien består av sju semistrukturerade intervjuer med kvinnor i åldrarna 16-25 som berättar om sina vanor på det sociala mediet Instagram. De sju kvinnorna visades 15 Instagrambilder från fem stora modeföretag och fick uttrycka sina spontana känslor inför dessa, likt ifall bilderna hade funnits i intervjupersonernas Instagramflöde. Bilderna bestod av två budskap som var uppdelade på en femgradig skala, där skalan rangordnades efter kategorierna extremt rationellt, rationellt, mixat, emotionellt och extremt emotionellt budskap. Intervjupersonerna kommenterade även vilka preferenser de hade angående bildernas innehåll ifall de själva fick välja. Efter att ha analyserat alla intervjuer menar vi att både de rationella och de emotionella bildkategorierna är att föredra, de står ut på den femgradiga skalan som både mer tilltalande och mer skapande av köplust. De extrema kategorierna av båda budskapen passar inte marknadsföring på Instagram av olika skäl. Med den här informationen hoppas vi kan hjälpa företag med deras marknadsföringsstrategi på Instagram.

Syftet med vår rapport är att undersöka huruvida humor förekommer i kosmetikareklamfilm på YouTube. Om humor nyttjas vill vi undersöka vilken humortyp som används. Vi vill också undersöka vilken berättarteknik som nyttjas samt vilken typ av reklambudskap som används. Vidare avser vi utreda huruvida humor är uppskattad i reklamfilm för kosmetika utifrån konsumentens perspektiv i Sverige. För att kunna undersöka detta används en visuell innehållsanalys till hjälp för att se hur kosmetikareklamfilmerna är uppbyggda. För att få en bild över hur konsumenten uppfattar kosmetikareklamfilm skapas en enkätundersökning. Där efterfrågas respondenternas allmänna uppfattning kring humor i kosmetikareklam samt deras åsikter om reklamfilmer som visas i enkäten. Undersökningen visar att humor inte används i någon större utsträckning i de kosmetikareklamfilmer som studeras. De flesta reklamfilmerna använder samma reklambudskap och berättarteknik. I de enstaka fall som humor förekom används endast tre humortyper av de sju som finns med i studien. Respondenterna uppskattar reklamfilmer med humorinslag då dessa upplevs mer positiva och personliga. / The purpose of our thesis is to investigate whether humor appears in cosmetic commercials on YouTube. If humor is implemented we would want to analyse which humor type it is. We also want to investigate which storytelling technique is used and what type of advertising message is applied. Furthermore, we intend to investigate whether humor is appropriate in commercials for cosmetics from the consumer's perspective in Sweden. In order to investigate this, a visual content analysis is used to help us see how the cosmetic commercials are constructed. A survey is created to get a picture of how the consumer perceives cosmetic commercials. The respondents are asked about their opinion on general perception about humor in cosmetics advertising and their opinion on commercials that are shown in the survey. The study shows that humor is not used to any great extent in the cosmetic commercials that are studied. Most of the commercials use the same advertising message and storytelling technique. In the few cases where humor occurres, only three types of humor are used of the seven included in the study. The respondents appreciate commercials with humor as they are seen more positive and personal.

Under det senaste årtiondet har djurrättsintresset i världen ökat markant. Samtidigt blir marknadsföring mer och mer viktigt i en allt mer digitaliserad värld, så hur marknadsför sig egentligen en djurrättsorganisation?  Syftet med föreliggande studie är att se hur en djurrättskampanj skulle kunna se ut generellt, genom att göra en visuell innehållsanalys på 20 utvalda reklamkampanjer för djurrättsorganisationer. För var reklamkampanj analyserades sex stycken reklamenheter. Resultatet presenterades sedan i form av tabeller och diagram, men även genom en egenskapad generisk reklamkampanj.   För att validera att den egenskapade reklamkampanjen uppfattas som generell användes en fokusgrupp där de olika enheternas generalitet diskuterades. Utefter detta justerades utformningen av bildspråket.  Undersökningen visade att det finns en del olikheter mellan olika djurrättsorganisationers reklamkampanjer, men även gemensamma nämnare kunde hittas. / During the last decade the interests in animal rights have increased significant. Meanwhile advertising becomes more and more important in a digitized world, so how do animal right organizations advertise themselves?  The purpose of this study is to look further into how a general animal rights campaign could be modeled, by implementing a visual content analysis on 20 selected advertising campaigns of animal right organizations. For each advertising campaign were six different advertising units analyzed. The result was later presented in table and diagram form, but also through a self-made generic advertising campaign.   To validate that the self-made advertising campaign is perceived as a generic one, we used focus groups where the generality of the different units was discussed. Out of this the modelling of the imagery was adjusted.  The study showed that there are some differences between the different animal right organization campaigns, but it also showed that common denominators can be found.