<p>Buildings are one of the fundamental sources of geospatial information for urban planning, population estimation, and infrastructure management. Although building extraction research has gained considerable progress through neural network methods, the labeling of training data still requires manual operations which are time-consuming and labor-intensive. Aiming to improve this process, this thesis developed an automated building extraction method based on the boundary following technique and the Mask Regional Convolutional Neural Network (Mask R-CNN) model. First, assisted by known building footprints, a boundary following method was used to automatically best label the training image datasets. In the next step, the Mask R-CNN model was trained with the labeling results and then applied to building extraction. Experiments with datasets of urban areas of Bloomington and Indianapolis with 2016 high resolution aerial images verified the effectiveness of the proposed approach. With the help of existing building footprints, the automatic labeling process took only five seconds for a 500*500 pixel image without human interaction. A 0.951 intersection over union (IoU) between the labeled mask and the ground truth was achieved due to the high quality of the automatic labeling step. In the training process, the Resnet50 network and the feature pyramid network (FPN) were adopted for feature extraction. The region proposal network (RPN) then was trained end-to-end to create region proposals. The performance of the proposed approach was evaluated in terms of building detection and mask segmentation in the two datasets. The building detection results of 40 test tiles respectively in Bloomington and Indianapolis showed that the Mask R-CNN model achieved 0.951 and 0.968 F1-scores. In addition, 84.2% of the newly built buildings in the Indianapolis dataset were successfully detected. According to the segmentation results on these two datasets, the Mask R-CNN model achieved the mean pixel accuracy (MPA) of 92% and 88%, respectively for Bloomington and Indianapolis. It was found that the performance of the mask segmentation and contour extraction became less satisfactory as the building shapes and roofs became more complex. It is expected that the method developed in this thesis can be adapted for large-scale use under varying urban setups.</p>

Syftet med studien var att studera hur män och kvinnor i 25-30 årsåldern reagerar på olika brottsanklagelser som sker på sociala medier. Undersökningen skedde genom semistrukturerade intervjuer med fem individer. Resultatet visade att respondenterna utan egen erfarenhet av att skriva ut misstänkta gärningspersoners namn var negativa till detta, samtidigt som respondenten som har skrivit ut brottsanklagelser ansåg att det var godtagbart. Resultaten visade även att brottstypen hade en viss påverkan gällande åsikterna, när det gällde brott som uppfattades som mindre var det oacceptabelt att publicera namnet, medan det noterades en större acceptans för grövre brott som exempelvis sexualbrott. Slutligen visade respondenternas svaratt trovärdigheten varierade beroende på vem det är som publicerat inlägget och hur den har uttryckt sig. Alla resultat tolkades i förhållande till stämplingsteorin och radikalfeministisk teori för att få en djupare förståelse för eventuella likheter och skillnader i respondenternas svar. / The purpose of this study was to examine how men and women aged 25-30 respond to various criminal accusations in social media. To investigate this, semi-structured interviews was conducted with five individuals. The results shows that the participants without own experience of posting accusations online were negative to post names of suspected perpetrators, while the participant with own experience considered it acceptable. Furthermore, it showed that the type of offense had some influence on the opinions. When it came to crimes considered as minor, it was unacceptable to publish the name, while greater acceptance was noted for more serious crimes such as sexual crimes. Finally, the participants based the credibility of a post on who published it and how they expressed themselves. All results were interpreted in relation to labeling theory and radical feminist theory to get a greater understanding of any similarities and differences in the participants answers.

This study examined the association between level of physical activity and changes in dietary behaviors of 3,945 employees after a 10-week work-site physical activity program. Fifty-seven percent of the participants met the CDC/ACSM standard for physical activity sufficient for a health benefit. Physical activity was not significantly related to increased fruit and vegetable consumption, decreased dietary fat and calorie intake, and participants acquiring new nutrition skills. Physical activity was negatively associated with increased food label awareness. Participants who exercised sufficient for a health benefit were less likely to increase their food label awareness. Physical activity and dietary behaviors are generally not associated. Interventions to improve these behaviors should be behavior-specific.

No description available.

The joining of two molecular functions or the strategy of adding functions to proteins has been tremendously important for the development of proteins as tools in research and clinic. Depending on the intended application, there are a wide variety of functions that can be added to a proteins. In clinical applications drugs are a commonly conjugated to antibodies and in research adding reporter groups such as biotin, enzymes or fluorophores is a routine procedure. The chemistries and methods most often used suffer from drawbacks such as lack of stringency, which could lead to undesired effects on the protein. Many site-specific methods of labeling of antibodies require modification or insertion of handles in the antibody recombinantly, before labeling can be performed. The core of this thesis is the development of a strategy for covalent specific labeling of antibodies by exploiting the site specific binding of the Z domain to Protein A. Photoreactive Z-domains were produced by solid phase peptide synthesis, which provides the opportunity to insert a photoreactive amino acid and a reporter biotin at specific positions in the domain. The inherited binding to the Fc-part of the antibody in combination with the incorporated photoreactive amino acid, BPA, is used for site-specific interaction, and thereafter, covalent coupling to the antibody. The exposure with the appropriate wavelength of light enables the formation a covalent linkage between the Z domain and the antibody. The biotinylated photoactivable domains were subsequently used to site-specifically label a number of different types of antibodies, polyclonal rabbit IgG, monoclonal human IgG1 and monoclonal mouse IgG2a, and thereafter the antibodies was employed in a variation of applications. The photolabeling procedure of antibodies by the use of photoactivable Z domains has proven to be successful and could serve as a valuable tool in several applications. / QC 20120507

Self-driving vehicles has become a hot topic in today's industry during the past years and companies all around the globe are attempting to solve the complex task of developing vehicles that can safely navigate roads and traffic without the assistance of a driver.  As deep learning and computer vision becomes more streamlined and with the possibility of using fisheye cameras as a cheap alternative to external sensors some companies have begun researching the possibility for assisted driving on vehicles such as electrical scooters to prevent injuries and accidents by detecting dangerous situations as well as promoting a sustainable infrastructure. However training such a model requires gathering large amounts of data which needs to be labeled by a human annotator. This process is expensive, time consuming, and requires extensive quality checking which can be difficult for companies to afford. This thesis presents an application that allows for semi-automatically labeling a dataset with the help of a human annotator and an object detector. The application trains an object detector together with an active learning framework on a small part of labeled data sampled from the woodscape dataset of fisheye distorted images and uses the knowledge of the trained model as well as using a human annotator as assistance to label more data. This thesis examines the labeled data produced by using the application described in this thesis and compares them with the quality of the annotations in the woodscape dataset. Results show that the model can't make any quality annotations compared to the woodscape dataset and resulted in the human annotator having to label all of the data, and the model achieved an accuracy of 0.00099 mAP.

This thesis deals with the attitudes of students of two secondary schools to marihuana, alcohol use and their users. Attitudes to alcohol, marihuana and to their users are topics frequently discussed in connection with changing perception of these two substances. This trend can be best seen in adolescents during the period of late adolescence which is the period of secondary school studies. The main aim of the thesis is to find out which factors influence the attitudes of students to alcohol and drug use and if there are any differences concerning those factors between the observed schools. These factors come from theoretical considerations which are subsequently tested in research. As the research showed, the students' attitudes to marihuana and alcohol use, and also to alcohol users are primarily influenced by significant others, among which belong mainly peers and their parents or their own experience. While the individuals form their attitudes to alcohol primarily according to their parents, in case of marihuana, the individuals, who have experience with it, follow mainly their peers and their own experience. Those individuals also tend to be more positive towards other marihuana users. There have not been found out any significant differences concerning the attitudes to marihuana and alcohol...

Research related to profiling and the outcome of traffic stops has generally focused on the race of the individuals involved. Little research has examined other characteristics, such as age and socioeconomic status, that may also play a role in traffic stop outcomes. The current study sought to address this limitation in two ways: (1) determine whether the characteristics of age, sex, race, social class, and demeanor are profiled during traffic stops and (2) whether these characteristics influenced the outcome of the traffic stops with regard to tickets and vehicle searches. Secondary data were utilized from the 2015 Police-Public Contact Survey. Findings revealed that not only race, but age, sex, social class, and demeanor of both the officer and the driver had an affect on the outcome of a traffic stops.

No description available.

The digastric muscle of the rabbit consists of a single anterior belly which inserts onto the lower jaw. Horseradish peroxidase was injected into the muscle and into subcutaneous regions overlying the lower jaw to determine the sites of origin of the motor innervation to both the digastric muscle and the platysma muscles. After digastric muscle injection, labelled cells were found in the ipsilateral retrotrigeminal nucleus as well as in the intermediate subnucleus of the main facial nucleus on both sides. Subcutaneous injections produced labelling which was found bilaterally in the intermediate subnucleus and in the ventromedial portion of the medial subnucleus. These results are interpreted in relation to the common embryological origin of these two muscles and their innervation.