We have developed a system to hunt and reuse special gene integration sites that allow for high and stable gene expression. A vector, named pRGFP8, was constructed. The plasmid pRGFP8 contains a reporter gene, gfp2 and two extraneous DNA fragments. The gene gfp2 makes it possible to screen the high expression regions on the chromosome. The extraneous DNA fragments can help to create the unique loci on the chromosome and increase the gene targeting frequency by increasing the homology. After transfection into Chinese hamster ovary cells (CHO) cells, the linearized pRGFP8 can integrate into the chromosome of the host cells and form the unique sites. With FACS, 90 millions transfected cells were sorted and the cells with strongest GFP expression were isolated, and then 8 stable high expression GFP CHO cell lines were selected as candidates for the new host cell. Taking the unique site created by pRGFP8 on the chromosome in the new host cells as a targeting locus, the gfp2 gene was replaced with the gene of interest, human ifngamma, by transfecting the targeting plasmid pRIH-IFN. Then using FACS, the cells with the dimmest GFP fluorescence were selected. These cells showed they had strong abilities to produce the protein of interest, IFN-gamma. During the gene targeting experiment, we found there is positive correlation between the fluorescence density of the GFP CHO host cells and the specific production rate of IFN-gamma. This result shows that the strategy in our expression system is correct: the production of the interesting protein increases with the increase fluorescence of the GFP host cells. This system, the new host cell lines and the targeting vector, can be utilized for highly expressing the gene of interest. More importantly, by using FACS, we can fully screen all the transfected cells, which can reduce the chances of losing the best cells. / Singapore-MIT Alliance (SMA)

Semiconductor scaling has been largely driven by advancements in lithographic technologies. However, the lack of a viable next generation lithography solution since the 180nm node has driven the industry to printing sub-wavelength features. This has led to rising manufacturing costs and diminishing chip yield. In traditional methodologies, manufacturing and design are relatively insulated, with a layout being the only means of communicating design intent to the foundry. In this dissertation, we describe several techniques which utilize electrical information to improve properties of manufactured structures. We aim to show that a bi-directional flow of information between design and manufacturing is key to increasing chip yield. In particular, we target the mask data preparation flow of lithography. We develop an electrically-driven optical proximity correction (ED-OPC) tool that performs electrical matching as opposed to geometric matching in order to achieve lower post-lithography delay error. We then demonstrate how to harness ED-OPC to compensate electrical variability arising from non-lithographic sources as well. We then describe a technique to manufacture circuits with less timing violations across the process window by using design-aware shape tolerances generated from timing information. Finally, we observe that local wiring has low impact on design properties and use this information to modify target wire shapes themselves in order to improve their manufacturability in the presence of process variations. / text

This thesis tackles the problem of automatically synthesizing motions of close-character interactions which appear in animations of wrestling and dancing. Designing such motions is a daunting task even for experienced animators as the close contacts between the characters can easily result in collisions or penetrations of the body segments. The main problem lies in the conventional representation of the character states that is based on the joint angles or the joint positions. As the relationships between the body segments are not encoded in such a representation, the path-planning for valid motions to switch from one posture to another requires intense random sampling and collision detection in the state-space. In order to tackle this problem, we consider to represent the status of the characters using the spatial relationship of the characters. Describing the scene using the spatial relationships can ease users and animators to analyze the scene and synthesize close interactions of characters. We first propose a method to encode the relationship of the body segments by using the Gauss Linking Integral (GLI), which is a value that specifies how much the body segments are winded around each other. We present how it can be applied for content-based retrieval of motion data of close interactions, and also for synthesis of close character interactions. Next, we propose a representation called Interaction Mesh, which is a volumetric mesh composed of points located at the joint position of the characters and vertices of the environment. This raw representation is more general compared to the tangle-based representation as it can describe interactions that do not involve any tangling nor contacts. We describe how it can be applied for motion editing and retargeting of close character interaction while avoiding penetration and pass-throughs of the body segments. The application of our research is not limited to computer animation but also to robotics, where making robots conduct complex tasks such as tangling, wrapping, holding and knotting are essential to let them assist humans for the daily life.

Submitted by CARLOS EDUARDO ROSSI CUBAS DA SILVA null (carlos.cubas@gmail.com) on 2017-02-25T17:53:14Z No. of bitstreams: 1 Dissertação-versão-final.pdf: 44412755 bytes, checksum: 4a1da7985f55f9ba238e1107610b8840 (MD5) / Approved for entry into archive by LUIZA DE MENEZES ROMANETTO (luizamenezes@reitoria.unesp.br) on 2017-03-07T13:42:27Z (GMT) No. of bitstreams: 1 silva_cerc_me_sjrp.pdf: 44412755 bytes, checksum: 4a1da7985f55f9ba238e1107610b8840 (MD5) / Made available in DSpace on 2017-03-07T13:42:27Z (GMT). No. of bitstreams: 1 silva_cerc_me_sjrp.pdf: 44412755 bytes, checksum: 4a1da7985f55f9ba238e1107610b8840 (MD5) Previous issue date: 2017-02-02 / Nas últimas décadas, o interesse quanto à captura de movimentos da face humana e à identificação de suas expressões com a finalidade de geração de animações faciais realistas, tem aumentado, tanto na comunidade científica quanto na indústria do entretenimento. A alta acurácia nesse processo é necessária, pois os humanos são treinados para identificar expressões faciais, detectando facilmente pequenas imperfeições na animação de uma face virtual. A animação facial baseada em performance é uma das técnicas utilizadas para gerar animações realistas principalmente em filmes. Com o surgimento de câmeras RGB-D de baixo custo, muitos sistemas de animação facial baseada em performance foram desenvolvidos, compartilhando muitos princípios fundamentais, mas com detalhes de implementação específicos. Estes sistemas consistem de uma fase de rastreamento de movimentos e identificação de expressões faciais seguido de um procedimento de redirecionamento de expressões. Neste sentido, ambientes modulares para a animação facial baseada em performance são extremamente úteis na incorporação de novos algoritmos de rastreamento com características de entrada e saídas padronizadas. Considerando o contexto exposto, o presente trabalho teve como principal objetivo a criação e validação de um ambiente com arquitetura modular para animação facial baseada em performance que utilizou uma câmera RGB-D para a captura dos movimentos faciais de um ator, bem como permitiu a incorporação dos principais algoritmos de rastreamento encontrados na literatura, visando o redirecionamento destes movimentos para uma face virtual humana diferente. As entradas e saídas deste ambiente foram padronizadas pelo uso de blendshapes (mistura de formas). / In recent decades, interest in capturing human face movements and identifying expressions for the purpose of generating realistic facial animations has increased in both the scientific community and the entertainment industry. The high accuracy in this process is necessary because humans are trained to identify facial expressions, easily detecting small imperfections in the animation of a virtual face. Performance-based facial animation is one of the techniques used to generate realistic animations especially in movies. With the emergence of low-cost RGB-D cameras, many performance-based facial animation systems have been developed, sharing many fundamental principles but with specific implementation details. These systems consist of a phase of tracking movements and identifying facial expressions followed by an expression redirection procedure. In this sense, modular environments for performance-based facial animation are extremely useful in incorporating new tracking algorithms with standardized input and output characteristics. Considering the above context, the main objective of this work was the creation and validation of an environment with modular architecture for performance-based facial animation that used an RGB-D camera to capture the facial movements of an actor as well as allowed the incorporation of the main algorithms found in the literary, aiming the redirection of these movements to a different human virtual face. The inputs and outputs of this environment were standardized by the use of blendshapes.

Attention retargeting är ett relativt ungt forskningsområde som bygger på forskningen om visuell prominens. Forskningen har hittills fokuserat på fotografier eller filmer och har inte applicerats på interaktiva medier såsom spel. Arbetet använde en förenklad metod av Mateescu och Bajíc (2014) där man justerar nyansen av intressepunkten 180◦ i en CIE Lab rymd. Metoden applicerades på fem tänkta spelrelaterade bilder. Arbetet genomfördes via kvalitativ intervjuer, inkluderande ögonstudier, med sju respondenter med olika spelvana. Resultatet visade att attention retargeting gav en högre prominens i de olika områdena där respondenterna tenderade att agera efter affordance mot de manipulerade objekten. Däremot så krävdes ett spelekologiskt perspektiv som respondenten kopplar till genom kunskap. När respondentens kunskap gjorde handlingen möjlig såg de kopplingen till objektet. Annars uteslöt respondenten handlingen för att finna en annan möjlig handling. Arbetet uppmanar till fortsatta studier där olika egenskaper testas. I framtiden skulle det ultimata målet kunna vara en subliminal attention retargeting där man styr en spelare i en till synes oförändrad miljö.

Titel: Retargeting is like a boomerang, what you give is what you get Forskningsfråga:Vilken inverkan har företags användning av retargeting på konsumenters köpintention? Vad är konsumenters attityd till vad som anses etiskt accepterat gällande retargeting? Syfte:Studiens syfte är att klarlägga hur konsumenter ställer sig till individanpassad marknadsföring genom cookies, så kallad retargeting. Studien syftar även till att belysa den inställning som existerar gällande de etiska aspekterna marknadsföringsverktyget avser och hur det påverkar konsumentens köpintention. Metod: Studien bygger på en deduktiv forskningsansats där forskarna utgått från befintlig teori och format testbara hypoteser. Hypoteserna har sedan behandlats genom en kvantitativ forskningsmetod med fokus på enkätundersökningar. Den data som samlats in har analyserats främst utifrån ett positivistiskt synsätt, med visst inslag av ett hermeneutistiskt synsätt. Studiens urval består av svenska konsumenter som infinner sig på internet. Resultat och slutsats: Retargeting är ett vanligt förekommande marknadsföringsverktyg som konsumenter dagligen exponeras för. Däremot kan retargeting idag upplevas som något mättat med hänsyn till att det upplevs som ett vanligt förekommande fenomen. Cookies är ett av flera verktyg som retargeting bygger på. Företag använder cookies i syfte att rikta sina annonser med relevant innehåll till lämpliga konsumenter. Studiens resultat har visat på att uppfattning och attityd är de två aspekter som påverkar köpintentionen rörande området. Kännedom och integritetsbekymmer som studien avsåg att undersöka visade sig inte vara de främsta aspekterna i relation till köpintention rörande retargeting. Avslutningsvis kan förevisa att såsom retargeting exploateras idag uppger konsumenter inneha en mer negativ inställning till marknadsföringen än positiv, vilket företag bör ta i beaktning. / Title: Retargeting is like a boomerang, what you give is what you get Research question: How does the influence of retargeting impact consumers purchase intentions? What is the consumers attitude towards what is considered to be ethically correct regarding retargeting? Purpose: The purpose of the research is to clarify how consumer respond to individualized marketing through cookies, also called retargeting. The study also put emphasis on the attitude that exist regarding the ethical aspect of the marketing tool and how it impact consumers purchase intentions. Method: The study builds upon a deductive research approach where the researchers have used an already established theory in order to develop further hypotheses. The hypotheses have been developed through quantitative research method with emphasis on surveys. The data collected have been analyzed through a positivistic viewpoint, with certain elements of hermeneutical viewpoint. The study consists of swedish consumers on the internet. Results and conclusions:Retargeting is a common marketing tool that consumers are facing on a daily basis. However, retargeting can appear saturated with regards to recurring phenomenon. Cookies is one out of many tools that retargeting builds upon. Businesses are using cookies for the purpose to target their ads with relevance to content suited to the expected consumer. The result have shown that interpretation and attitude are the two major aspects that influence purchase intention. Knowledge and privacy violations were not two of the major aspects that would influence the purchase intention regarding retargeting. Consequently, the use of retargeting have resulted in a significant negative attitude among consumers, which businesses should highly take into consideration.

The term cookie regarding IT is still a relatively new term. It is rarely discussed among internet users. The name cookie is fairly known but the concept not as much. Cookies are well established by different companies and organizations that in some way offer online services. Today visitors of the internet meet cookies almost every time they go online. For the users who choose not to accept cookies the online experience becomes highly limited. Cookies store information of the visitors in order to improve their experience but also to help the organization. The majority of webpages online that uses electronic commerce or in any way can benefit from storing the customers information uses cookies. Cookies can be sold or exchanged between organizations to increase their profit and range. In 2011 a provision regarding all organizations using cookies were implemented. Every internet user visiting these pages had to be notified about their use of cookies immediately. New laws are established in order to control or regulate the exertion of other people’s personal information at the same pace as cookies are further developed. The next law to be established is GDPR, General Data Protection Regulation, which will be focused on in this thesis. The purpose of this study is to explore the phenomenon of cookie-use and how it affects internet users and organizations in synergy with the GDPR-law. To achieve the desired result a mixed method has been applied involving both interviews and a survey. The interviews consisted of pre-determined questions and dynamic follow-up questions. Three different organizations participated in the interviews, an university, an organization using e-business and a media agency. The survey consisted of a few mandatory closed questions. The distribution of the survey led to right above 130 respondents. The gained results from the use of a quantitative methods hinted that barely half of the respondents knew what a cookie actually was. Several of the participants were curious about learning more about the phenomenon while a lower percentage did not consider it important. Knowledge or ignorance regarding cookies turned out to have a great impact on the respondent’s comfortableness and behavior online. The results from the qualitative method showed that the organizations working with cookies will be affected by GDPR. Following thesis will contribute to the field of informatics by consequently process the empiric data collected with support from theoretic competences. It will contribute as a source of information regarding cookies within information technology. / Termen cookie är fortfarande en relativt ny term och är inte särskilt debatterad bland internetanvändare. Begreppet cookie är någorlunda känt men innebörden inte lika så. Cookies är väl etablerat av olika företag och organisationer som på något sätt använder sig av internet. Idag stöter användare på cookies nästan överallt när de rör sig på internet. För den som bestämmer sig för att avstå från allt som har med cookies att göra blir internet vistelsen mycket begränsad. Cookies arbetar för att lagra information om besökarna dels för att gynna besökarnas vistelse men även för att hjälpa företagen. Majoriteten av alla hemsidor som idag använder sig av e-handel samt kan gynnas av att kundens data sparas använder sig av cookies. Cookies kan även säljas eller utbytas mellan företag och organisationer för att öka deras försäljning och vidd. År 2011 implementerades en bestämmelse gällande alla företag och organisationer som då använde sig av cookies. Alla besökare skulle upplysas om att just de använde cookies när de besöker deras hemsida. Nya lagar etableras för att kunna kontrollera och reglera användandet av besökares personliga information online i takt med att användandet av cookie utvecklas. Den nästkommande lagen att implementeras är GDRP, General Data Protection Regulation, som kommer att ha stort fokus i denna uppsats. Syftet med denna uppsats är att granska fenomenet cookie, dess användningsområden samt hur det påverkar internetanvändare och de olika företagen och organisationerna i synergi med GDPR-lagen. För att uppnå önskat resultat har en mixad metod använts i form av intervjuer och en enkät. Intervjuerna bestod av förutbestämda frågor och dynamiska följdfrågor under samtalets gång. Tre företag ställde upp som intervjuobjekt. En högskola, ett företag med försäljning online samt en mediebyrå. Enkäten bestod av obligatoriska frågor med förutbestämda svarsalternativ. Distributionen av enkäten ledde till strax över 130 respondenter. Resultatet efter användningen av den kvantitativa metoden antydde att knappt hälften av respondenterna inte visste vad en cookie var. Många var nyfikna och ville lära sig mer medan en lägre procent av respondenterna inte ansåg att de inte ville veta mer. Vetskapen samt ovetskapen gällande cookies påvisade sig ha stor inverkan på respondenternas bekvämlighet samt beteende online. Resultatet av den kvalitativa metoden visade att organisationer som arbetar med cookies kommer bli påverkade av GDPR. Följande uppsats kommer bidra till området informatik genom att följaktligen bearbeta empiriinsamling med stöd av teoretiska kompetenser. Den kommer att bidra som en informationskälla gällande cookies inom informationsteknologi.

The diversity of video material, where a video is shot and produced using a single aspect ratio, and the variety of devices that can play video with screens in different aspect ratios make video retargeting a relevant topic. The process of fitting a video filmed in one aspect ratio to a screen in another aspect ratio is called video retargeting, and the retargeted video should ideally preserve the important content and structure of the original video as well as be free of visual artifacts. Important content and important structure are vague and subjective definitions, which makes this problem more difficult to solve. The video retargeting problem has been a challenge for researchers from the computer vision, computer graphics and human-computer interaction areas, and successful retargeting can improve the viewing experience and the content’s aesthetic value. Video retargeting is done by four tools: cropping, scaling, seam carving and seam adding. Previous research showed that one of the keys to successful retargeting is to use a suitable combination of operators. This study makes use of a vision transformer, a deep learning model which is trained to discriminate between original and retargeted videos. Solving an optimization problem using beam search, the transformer assists in choosing a combination of operators that will result in the best possible retargeted video. The retargeted videos were examined in a user A/B-test, where users had to choose their preferred variant of a video shot: the transformer’s output using beam search, or a singular version where the video underwent a single retargeting operation. The model and user preferences were compared to check if the model indeed can make retargeting decisions that are appealing for humans to watch. A significance test showed that no conclusion can be made, probably due to lack of enough test data. However, the study revealed patterns in the preferences of the users and the model that could be further fine-tuned or combined with other computer vision mechanisms in order to output better retargeted videos. / Variation av videomaterial, där olika videor är inspelade och producerade i olika bildförhållande, samt variation i apparater och skärmar som spelar upp videor i olika bildförhållanden gör ändring av videobildförhållande till en relevant fråga. Processen där en videos bildförhållande ändras heter video retargeting. När video retargeting används bör den nya videon helst bevara strukturen och viktigt innehåll från originalvideon samt vara artefaktfri. Struktur och viktigt innehåll är subjektiva definitioner vilket gör frågan svårlöst, och frågan har varit en utmaning för forskare inom datorseende, datorgrafik och människa-datorinteraktion. Lyckad ändring av en videos bildförhållande kan förbättra tittarupplevelsen och innehållets estetiska värde. Video retargeting kan göras med hjälp av fyra funktioner: klippning, skalning, seam carving och seam adding. Tidigare studier visar att en av nycklarna till lyckad retargeting är att hitta en lämplig kombination av funktionerna. I denna studie används Vision Transformer, en djupinlärningsmodell som tränas för att skilja mellan original och omvandlade videor. Genom att lösa ett optimeringsproblem med strålsökning hjälper modellen välja den kombination av funktionerna som resulterar i den bästa möjliga omvandlade videon. De omvandlade videorna testades genom ett användartest där användare valde vilket videoklipp de tyckte bättre om: modellens output som skapades med hjälp av strålsökning, eller en version där klippet genomgick en enklare retargeting med hjälp av endast en av funktionerna. Modellens och användarnas preferenser jämfördes för att se om modellen kan fatta beslut som användare upplever som bra. Ett signifikanstest visar att ingen slutsats kan dras, förmodligen på grund av det begränsade antalet videoklipp och data som användes i studien. Däremot visar studien mönster i användarnas och modellens preferenser som kan användas för att vidareutveckla problemlösningen inom området.

In this thesis, we address a variety of challenges for analysis and enhancement of Computational Video. We present novel post-processing methods to bridge the difference between professional and casually shot videos mostly seen on online sites. Our research presents solutions to three well-defined problems: (1) Video stabilization and rolling shutter removal in casually-shot, uncalibrated videos; (2) Content-aware video retargeting; and (3) spatio-temporal video segmentation to enable efficient video annotation. We showcase several real-world applications building on these techniques. We start by proposing a novel algorithm for video stabilization that generates stabilized videos by employing L1-optimal camera paths to remove undesirable motions. We compute camera paths that are optimally partitioned into constant, linear and parabolic segments mimicking the camera motions employed by professional cinematographers. To achieve this, we propose a linear programming framework to minimize the first, second, and third derivatives of the resulting camera path. Our method allows for video stabilization beyond conventional filtering, that only suppresses high frequency jitter. An additional challenge in videos shot from mobile phones are rolling shutter distortions. Modern CMOS cameras capture the frame one scanline at a time, which results in non-rigid image distortions such as shear and wobble. We propose a solution based on a novel mixture model of homographies parametrized by scanline blocks to correct these rolling shutter distortions. Our method does not rely on a-priori knowledge of the readout time nor requires prior camera calibration. Our novel video stabilization and calibration free rolling shutter removal have been deployed on YouTube where they have successfully stabilized millions of videos. We also discuss several extensions to the stabilization algorithm and present technical details behind the widely used YouTube Video Stabilizer. We address the challenge of changing the aspect ratio of videos, by proposing algorithms that retarget videos to fit the form factor of a given device without stretching or letter-boxing. Our approaches use all of the screen's pixels, while striving to deliver as much video-content of the original as possible. First, we introduce a new algorithm that uses discontinuous seam-carving in both space and time for resizing videos. Our algorithm relies on a novel appearance-based temporal coherence formulation that allows for frame-by-frame processing and results in temporally discontinuous seams, as opposed to geometrically smooth and continuous seams. Second, we present a technique, that builds on the above mentioned video stabilization approach. We effectively automate classical pan and scan techniques by smoothly guiding a virtual crop window via saliency constraints. Finally, we introduce an efficient and scalable technique for spatio-temporal segmentation of long video sequences using a hierarchical graph-based algorithm. We begin by over-segmenting a volumetric video graph into space-time regions grouped by appearance. We then construct a "region graph" over the obtained segmentation and iteratively repeat this process over multiple levels to create a tree of spatio-temporal segmentations. This hierarchical approach generates high quality segmentations, and allows subsequent applications to choose from varying levels of granularity. We demonstrate the use of spatio-temporal segmentation as users interact with the video, enabling efficient annotation of objects within the video.

Robust manipulation capture and retargeting has been a longstanding goal in both the fields of animation and robotics. In this thesis I describe a new approach to capture both the geometry and motion of interactions with objects, dealing with the problems of occlusion by the use of magnetic systems, and performing the reconstruction of the geometry by an RGB-D sensor alongside visual markers. This ‘interaction capture’ allows the scene to be described in terms of the spatial relationships between the character and the object using novel topological representations such as the Electric Parameters, which parametrise the outer space of an object using properties of the surface of the object. I describe the properties of these representations for motion generalisation and discuss how they can be applied to the problems of human-like motion generation and programming by demonstration. These generalised interactions are shown to be valid by demonstration of retargeting grasping and manipulation to robots with dissimilar kinematics and morphology using only local, gradient-based planning.