No description available.

Image augmentation is a field that covers the subject area of altering existing data to create more for the use of model training processes. It may be seen as the practice of expanding upon existing data using a range of techniques that employ transformations to improve the diversity of training sets when applied to machine learning. In our case of image recognition, triplet loss is utilised to pair a reference image to a matching and non-matching input. However, since there are many single images, augmentation techniques are relied upon to expand upon our data set to improve the recognition of images and create true positives. True positives are created using standard augmentation techniques like perspective transformation, contrast, cropping, and more. Despite this, the same images may undergo other types of alterations, natural disturbances such as transformations and warping, that are not captured by standard augmentation techniques. Such instances constitute to the variance in identification. Therefore, the analysis of augmentations by artificial intelligence (AI) based recognition is proposed; AI is used in order to identify what contributes to realistic disturbances of single images that better imitate real-life transformations. Analysing existing standard image augmentation techniques should provide further insight within this scope, as to better determine ways of emulating natural disturbances, and the formulation of non-standard practices in tandem. How this is done is by the use of an image's identity, the pixels it's comprised of and their distributions. Through a methodology of inspecting image identities, the breaking down of augmentations, and the inquiry into practices of non-standard image augmentation techniques, we detect the variance in accuracy of generated models, analysing the comprised data sets. Our results show that augmentations improve accuracy on a basis of variance and divergence from the original image. Subsequent discussion expands upon the identities of images and how augmentations must still resemble true positives, with the potential of an augmentation gauged by its influence on the rate of growth and highest accuracy of a model. / <p></p><p></p><p></p>

Traditional image processing solutions have been applied in the past to detect and count strawberries. These methods typically involve feature extraction followed by object detection using one or more features. Some object detection problems can be ambiguous as to what features are relevant and the solutions to many problems are only fully realized when the modern approach has been applied and tested, such as deep learning. In this work, we investigate the use of R-CNN for green strawberry detection. The object detection involves finding regions of interest (ROIs) in field images using the selective segmentation algorithm and inputting these regions into a pre-trained deep neural network (DNN) model. The convolutional neural networks VGG, MobileNet and ResNet were implemented to detect subtle differences between green strawberries and various background elements. Downscaling factors, intersection over union (IOU) thresholds and non-maxima suppression (NMS) values can be tweaked to increase recall and reduce false positives while data augmentation and negative hardminging can be used to increase the amount of input data. The state of the art model is sufficient in locating the green strawberries with an overall model accuracy of 74%. The R-CNN model can then be used for crop yield prediction to forecast the actual red strawberry count one week in advance with a 90% accuracy.

It has been shown through several research investigations that users tend to underestimate distances in virtual reality (VR). Virtual objects that appear close to users wearing a Head-mounted display (HMD) might be located at a farther distance in reality. This discrepancy between the actual distance and the distance observed by users in VR was found to hinder users from benefiting from the full in-VR immersive experience, and several efforts have been directed toward finding the causes and developing tools that mitigate this phenomenon. One hypothesis that stands out in the field of spatial perception is the rest frame hypothesis (RFH), which states that visual frames of reference (RFs), defined as fixed reference points of view in a virtual environment (VE), contribute to minimizing sensory mismatch. RFs have been shown to promote better eye-gaze stability and focus, reduce VR sickness, and improve visual search, along with other benefits. However, their effect on distance perception in VEs has not been evaluated. To explore and better understand the potential effects that RFs can have on distance perception in VR, we used a blind walking task to explore the effect of three head-centric RFs (a mesh mask, a nose, and a hat) on egocentric distance estimation. We performed a mixed-design study where we compared the effect of each of our chosen RFs across different environmental conditions and target distances in different 3D environments. We found that at near and mid-field distances, certain RFs can improve the user's distance estimation accuracy and reduce distance underestimation. Additionally, we found that participants judged distance more accurately in cluttered environments compared to uncluttered environments. Our findings show that the characteristics of the 3D environment are important in distance estimation-dependent tasks in VR and that the addition of head-centric RFs, a simple avatar augmentation method, can lead to meaningful improvements in distance judgments, user experience, and task performance in VR.

During the broadcast of a golf tournament, a way to show the audience what a player's swing looks like would be to draw a trace following the movement of the club head. A computer vision model can be trained to identify the position of the club head in an image, but due to the high speed at which professional players swing their clubs coupled with the low frame rate of a typical broadcast camera, the club head is not discernible whatsoever in most frames. This means that the computer vision model is only able to deliver a few sparse detections of the club head. This thesis project aims to develop a machine learning model that can predict the complete motion of the club head, in the form of a swing trace, based on the sparse club head detections. Slow motion videos of golf swings are collected, and the club head's position is annotated manually in each frame. From these annotations, relevant data to describe the club head's motion, such as position and time parameters, is extracted and used to train the machine learning models. The dataset contains 256 annotated swings of professional and competent amateur golfers. The two models that are implemented in this project are XGBoost and a feed forward neural network. The input given to the models only contains information in specific parts of the swing to mimic the pattern of the sparse detections. Both models learned the underlying physics of the golf swing, and the quality of the predicted traces depends heavily on the amount of information provided in the input. In order to produce good predictions with only the amount of input information that can be expected from the computer vision model, a lot more training data is required. The traces predicted by the neural network are significantly smoother and thus look more realistic than the predictions made by the XGBoost model.

Burst abdomen (BA) is a severe complication after abdominal surgery, which often requires urgent repair. However, evidence on surgical techniques to prevent burst abdomen recurrence (BAR) is scarce. We conducted a retrospective analysis of patients with BA comparing them to patients with superficial surgical site infections from the years 2015 to 2018. The data was retrieved from the institutional wound register. We analyzed risk factors for BA occurrence as well as its recurrence after BA repair and surgical closure techniques that would best prevent BAR.:1 Abkürzungsverzeichnis 2 Einführung 2.1 Aufbau der Bauchwand und operative Zugangswege in der Abdominalchirurgie 2.1.1 Anatomie 2.1.2 Zugangswege 2.2 Wundinfektionen 2.3 Definition „Platzbauch“ 2.4 Risikofaktoren und Ursachen von Fasziendehiszenzen 2.4.1 Biochemische Einflüsse auf die Wundheilung 2.4.2 Mechanische und technische Faktoren 2.4.3 Allgemeine individuelle Faktoren 2.5 Management des Platzbauchs 2.6 Spätkomplikationen des Platzbauches 2.6.1 Narbenhernien 2.6.2 Intestinale Fisteln 2.6.3 Netzinfektion 2.6.4 Re-Dehiszenzen 3 Zielsetzung der vorliegenden Arbeit 4 Publikation 5 Zusammenfassung der Arbeit 5.1 Einleitung 5.2 Wundregister nosokomialer Wundinfektionen der Klinik für Viszeral-, Transplantations-, Thorax- und Gefäßchirurgie am Universitätsklinikum Leipzig 5.3 Risikofaktoren für Platzbäuche 5.4 Platzbauchentstehung 5.5 Chirurgische Verschlusstechnik 5.6 Re-Dehiszenzen 5.7 Limitationen der Analyse 6 Literaturverzeichnis 7 Anlagen 7.1 Darstellung des eigenen Beitrags 7.2 Selbstständigkeitserklärung 7.3 Lebenslauf 7.4 Publikationen 8 Danksagung

Considerable effort has been devoted to finding reliable methods of correcting noisy RGB-D images captured with unreliable depth-sensing technologies. Supervised neural networks have been shown to be capable of RGB-D image correction, but require copious amounts of carefully-corrected ground-truth data to train effectively. Data collection is laborious and time-intensive, especially for large datasets, and generation of ground-truth training data tends to be subject to human error. It might be possible to train an effective method on a relatively smaller dataset using synthetically damaged depth-data as input to the network, but this requires some understanding of the latent noise distribution of the respective camera. It is possible to augment datasets to a certain degree using naive noise generation, such as random dropout or Gaussian noise, but these tend to generalize poorly to real data. A superior method would imitate real camera noise to damage input depth images realistically so that the network is able to learn to correct the appropriate depth-noise distribution.We propose a novel noise-generating CNN capable of producing realistic noise customized to a variety of different depth-noise distributions. In order to demonstrate the effects of synthetic augmentation, we also contribute a large novel RGB-D dataset captured with the Intel RealSense D415 and D435 depth cameras. This dataset pairs many examples of noisy depth images with automatically completed RGB-D images, which we use as proxy for ground-truth data. We further provide an automated depth-denoising pipeline which may be used to produce proxy ground-truth data for novel datasets. We train a modified sparse-to-dense depth-completion network on splits of varying size from our dataset to determine reasonable baselines for improvement. We determine through these tests that adding more noisy depth frames to each RGB-D image in the training set has a nearly identical impact on depth-completion training as gathering more ground-truth data. We leverage these findings to produce additional synthetic noisy depth images for each RGB-D image in our baseline training sets using our noise-generating CNN. Through use of our augmentation method, it is possible to achieve greater than 50% error reduction on supervised depth-completion training, even for small datasets.

Open Domain Question Answering (OpenQA) systems find an answer to a question from a large collection of unstructured documents. In this information era, we have an immense amount of data at our disposal. However, filtering all the content and trying to find the answers to our questions can be too time-consuming and ffdiicult. In addition, in such a globalised world, the information we look for to answer a question may be in a different language. Current research is focused on improving monolingual (English) OpenQA performance. This creates a disparity between the tools accessible between English and non-English speakers. The techniques explored in this study involve the combination of different methods, such as data augmentation and hard negative filtering for performance increase, and binary embeddings for improving the efficiency, with multilingual Transformers. The downstream performance is evaluated using sentiment multilingual datasets covering Cross-Lingual Transfer (XLT), question and answer in the same language, and Generalised Cross-Lingual Transfer (G-XLT), different languages for question and answer. The results show that data augmentation increased Recall by 37.0% and Mean Average Precision (MAP) by 67.0% using languages absent from the test set for XLT. Combining binary embeddings and hard negatives can reduce inference time and index size to 12.5% and 3.1% of the original, retaining 97.1% of the original Recall and 94.8% of MAP (averages of XLT and MAP). / Open Domain Question Answering (OpenQA)-system hittar svar på frågor till stora samlingar av ostrukturerade dokument. I denna informationsepok har vi en enorm mängd kunskap till vårt förfogande. Att filtrera allt innehåll för att försöka att hitta svar på våra frågor kan dock vara mycket tidskrävande och svårt. I en globaliserad värld kan informationen vi söker för att besvara en fråga dessutom vara på ett annat språk. Nuvarande forskning är primärt inriktad på att förbättra OpenQA:s enspråkiga (engelska) prestanda. Detta skapar ett gap mellan de verktyg som är tillgängliga för engelsktalande och icke-engelsktalande personer. De tekniker som undersöks i den här studien innebär en kombination av olika metoder, t.ex. dataförstärkning och hård negativ filtrering för att öka prestandan, och binära embeddings för att förbättra effektiviteten med flerspråkiga Transformatorer. Prestandan nedströms utvärderas med hjälp av flerspråkiga dataset som omfattar Cross-Lingual Transfer (XLT), fråga och svar på samma språk, och Generalised Cross-Lingual Transfer (G-XLT), olika språk för fråga och svar. Resultaten visar att dataförstärkning ökade recall med 37.0% och 67.0% för Mean Average Precision (MAP) med hjälp av språk som inte fanns med i testuppsättningen för XLT. Genom att kombinera binära embeddings och hårda negationer kan man minska tiden för inferens och indexstorleken till 12.5% och 3.1% av originalet, samtidigt som man behåller 97.1% av ursprunglig recall samt 94.8% av MAP (medelvärden av XLT och MAP).

No description available.

No description available.