This thesis presents detailed observation of different Convolutional Neural Network (CNN) architecture which had assisted Computer Vision researchers to achieve state-of-the-art performance on classification, detection, segmentation and much more to name image analysis challenges. Due to the advent of deep learning, CNN had been used in almost all the computer vision applications and that is why there is utter need to understand the miniature details of these feature extractors and find out their pros and cons of each feature extractor meticulously. In order to perform our experimentation, we decided to explore an object detection task using a particular model architecture which maintains a sweet spot between computational cost and accuracy. The model architecture which we had used is LSTM-Decoder. The model had been experimented with different CNN feature extractor and found their pros and cons in variant scenarios. The results which we had obtained on different datasets elucidates that CNN plays a major role in obtaining higher accuracy and we had also achieved a comparable state-of-the-art accuracy on Pedestrian Detection Dataset. In extension to object detection, we also implemented two different model architectures which find shoulder keypoints. So, One of our idea can be explicated as follows: using the detected annotation from object detection, a small cropped image is generated which would be feed into a small cascade network which was trained for detection of shoulder keypoints. The second strategy is to use the same object detection model and fine tune their weights to predict shoulder keypoints. Currently, we had generated our results for shoulder keypoint detection. However, this idea could be extended to full-body pose Estimation by modifying the cascaded network for pose estimation purpose and this had become an important topic of discussion for the future work of this thesis.

This thesis is all about the visual attention, starting from understanding the human visual system up till applying this mechanism to a real-world computer vision application. This has been achieved by taking the advantage of latest findings about the human visual attention and the increased performance of the computers. These two facts played a vital role in simulating the many different aspects of this visual behavior. In addition, the concept of bio-inspired visual attention systems have become applicable due to the emergence of different interdisciplinary approaches to vision which leads to a beneficial interaction between the scientists related to different fields. The problems of high complexities in computer vision lead to consider the visual attention paradigm to become a part of real time computer vision solutions which have increasing demand.  In this thesis work, different aspects of visual attention paradigm have been dealt ranging from the biological modeling to the real-world computer vision tasks implementation based on this visual behavior. The implementation of traffic signs detection and recognition system benefited from this mechanism is the central part of this thesis work.

This thesis is focused on the acceleration of Random Forest object detection in an image. Random Forest detector is an ensemble of independently evaluated random decision trees. This feature can be used to acceleration on graphics unit. Development and increasing performance of graphics processing units allow the use of GPU for general-purpose computing (GPGPU). The goal of this thesis is describe how to implement Random Forest method on GPU with OpenCL standard.

Bone age assessment can be useful in a variety of ways. It can help pediatricians predict growth, puberty entrance, identify diseases, and assess if a person lacking proper identification is a minor or not. It is a time-consuming process that is also prone to intra-observer variation, which can cause problems in many ways. This thesis attempts to improve and speed up bone age assessments by using different object detection methods to detect and segment bones anatomically important for the assessment and using these segmented bones to train deep learning models to predict bone age. A dataset consisting of 12811 X-ray hand images of persons ranging from infant age to 19 years of age was used. In the first research question, we compared the performance of three state-of-the-art object detection models: Mask R-CNN, Yolo, and RetinaNet. We chose the best performing model, Yolo, to segment all the growth plates in the phalanges of the dataset. We proceeded to train four different pre-trained models: Xception, InceptionV3, VGG19, and ResNet152, using both the segmented and unsegmented dataset and compared the performance. We achieved good results using both the unsegmented and segmented dataset, although the performance was slightly better using the unsegmented dataset. The analysis suggests that we might be able to achieve a higher accuracy using the segmented dataset by adding the detection of growth plates from the carpal bones, epiphysis, and the diaphysis. The best performing model was Xception, which achieved a mean average error of 1.007 years using the unsegmented dataset and 1.193 years using the segmented dataset. / <p>Presentationen gjordes online via Zoom. </p>

Cette thèse présente une tentative d'approche du problème de la détection et discrimination des petits véhicules dans des images aériennes en vue verticale par l'utilisation de techniques issues de l'apprentissage profond ou "deep-learning". Le caractère spécifique du problème permet d'utiliser des techniques originales mettant à profit les invariances des automobiles et autres avions vus du ciel.Nous commencerons par une étude systématique des détecteurs dits "single-shot", pour ensuite analyser l'apport des systèmes à plusieurs étages de décision sur les performances de détection. Enfin nous essayerons de résoudre le problème de l'adaptation de domaine à travers la génération de données synthétiques toujours plus réalistes, et son utilisation dans l'apprentissage de ces détecteurs. / The following manuscript is an attempt to tackle the problem of small vehicles detection in vertical aerial imagery through the use of deep learning algorithms. The specificities of the matter allows the use of innovative techniques leveraging the invariance and self similarities of automobiles/planes vehicles seen from the sky.We will start by a thorough study of single shot detectors. Building on that we will examine the effect of adding multiple stages to the detection decision process. Finally we will try to come to grips with the domain adaptation problem in detection through the generation of better looking synthetic data and its use in the training process of these detectors.

Inom arbetslivet finns situationer i vilka vi kan utnyttja digitalisering för att förenkla och effektivisera arbetet. Ett exempel är den analoga hanteringen av fysiska formulär. Oftast överförs data från fysiska formulär till datorn manuellt. Syftet med detta projekt är att effektivisera den generella hanteringen av pappersformulär genom inskanning. Detta kan göras genom att utnyttja en beskärningsfunktion vid inskanningen. Beskärningen används för att beskära bort irrelevant data från formuläret och därmed framhävs det som ska skannas in. Därefter kan objektigenkänning användas för att känna igen siffror och text från det framhävda fältet. En Androidapplikation har utvecklats som utnyttjar mobilens inbyggda kamera för att skanna in och framhäva viktiga fält från formulär. Parallellt tränades en maskininlärningsmodell, med TensorFlow, att känna igen handskrivna siffror. Den färdigtränade modellen jämfördes med olika OCR-verktyg och resultatet visade att modellen detekterar handskrivna siffror bättre. / A workplace can be made more efficient by digitalization. An example of that is the handling of forms. Most of the time physical forms are manually digitalized. The aim of this project is to simplify the general handling of forms by automating the process. This could be done by scanning photos of forms and using a cropping function to highlight the important parts. By doing this we can use object detection to recognize the text or numbers on that highlighted field. An application was built that utilizes a phone camera to snap a photo of a form, and then a cropping function was implemented to crop out the important part of the form excluding irrelevant data. Parallel to that a machine learning model was trained with TensorFlow to recognize handwritten numbers to work with the application. The trained model was evaluated and compared to different OCR tools, and the results showed that a model trained to detect a specific handwriting works better than general OCR tools on handwritten digits.

Motivet bakom examensarbetet är att skapa ett kameraövervakningssystem, ämnadför att filma och fotografera fiskgjuseföräldrarna, när de är vid deras avkommor ifiskgjusebon. Inspelningarna och bilderna ska användas för att studera fiskgjuseför-äldrarnas beteende vid deras avkommor. Litteraturstudien visar att fiskgjusen skaparsina bon på höga plattkronade tallar. Dessutom att människors närvaro vid fiskgjuse-bona har en negativ påverkan, då fiskgjuseungar minskar i antal. Likväl har forskareskapat kameraövervakning ämnad för att filma fågelbo. Kameraövervakningssyste-met kommer att medföra att människors närvaro minimeras vid fiskgjuseboet. Nu-förtiden används generellt Raspberry Pi, vilket är en dator byggd på ett litetkretskort. Raspberry Pi kan anslutas med en kamera och lagra inspelningar på SD-minneskortet. Dessutom går det att ansluta diverse sensorer, installera diversemjukvaror och programmeras. Metoden för studien är att skriva koder för att skapa program som göra att Rasp-berry Pi startar inspelningar och tar foton. Dessa program skrivs i Python, biblioteksom Open CV och COCO datablad används. Studien omfattar en konstruktion av ett kameraövervakningssystem, med ett Rasp-berry Pi, högkvalitets kamera, infrarödsensor, ljudsensor och program för att ta fo-ton och starta filmningen. Filmerna sparas på SD-minneskortet. Kameraövervak-ningssystemet strömkälla blir en power bank. Resultatet blev ett program som tar ett foto när sensorerna detektera antingen infra-röd strålning eller ljud eller förändringar i pixlar från kamerabilden. Fotot analyserasav programmet för att se om det är några fåglar på fotot. I programmet går det attställa in hur många fåglar som ska vara med på bilden för att den ska starta inspelningoch spara fotot. Sedan gjordes ett program som hela tiden känner av om det finnsfåglar på kamerabilden. När önskvärda antal fåglar är med på kamerabilden, börjarkameran spela in och tar sedan ett foto. / The motive behind the thesis is to create a camera surveillance system, intended forfilming and photographing the osprey parents, when they are with their offspring, inthe osprey nest. The recordings and pictures will be used to study the osprey par-ents' behavior, nearby their offspring. The literature study shows that ospreys createtheir nests on tall, flat-crowned pines. In addition, the presence of humans near os-prey nests has a negative impact, as osprey chicks decrease in number. Nevertheless,researchers have created camera surveillance intended for filming bird nests. Thecamera surveillance system will minimize human presence at the osprey nest. Now-adays, the Raspberry Pi is used, which is a computer built on a small circuit board.The Raspberry Pi can be connected with a camera and store recordings on the SDmemory card. In addition, it is possible to connect various sensors, install varioussoftware and program. The method of the study is to write codes to create programs that make the Rasp-berry Pi start recording and taking photos. These programs are written in Python,libraries such as Open CV and COCO datasheets are used. The study includes the design of a camera surveillance system, with a Raspberry Pi,a high-quality camera, an infrared sensor, a sound sensor, a program for taking pho-tos and to start filming. The recordings and photos are saved on the SD memorycard. The camera surveillance system is powered by a power bank. The result was a program that takes a photo when the sensors detect either infraredradiation or sound or changes in pixels from the camera image. The photo is ana-lyzed by the program to see if there are any birds in the photo. In the program it ispossible to set how many birds should be in the picture for it to start recording andsave the photo. Then a program was made that continuously detects if there arebirds in the camera image. When the required number of birds is in the camera im-age, the camera starts recording and then takes a photo.

500 of 690 training images used in optimized training runs.

Annotations for 500 of the 690 images used for training.

In recent years, deep learning based computer vision technology has developed rapidly. This is not only due to the improvement of computing power, but also due to the emergence of high-quality datasets. The combination of object detectors and drones has great potential in the field of rescue and disaster relief. We created an image dataset specifically for vision applications on drone platforms. The dataset contains 5000 images, and each image is carefully labeled according to the PASCAL VOC standard. This specific dataset will be very important for developing deep learning algorithms for drone applications. In object detection models, loss function plays a vital role. Considering the uneven distribution of large and small objects in the dataset, we propose adjustment coefficients based on the frequencies of objects of different sizes to adjust the loss function, and finally improve the accuracy of the model.