Master of Science / Department of Electrical and Computer Engineering / Christopher L. Lewis / This thesis presents a novel approach for detecting robust and scale invariant interest points in images. The detector accurately and efficiently approximates the Laplacian of Gaussian using an optimal set of weighted box filters that take advantage of integral images to reduce computations. When combined with state-of-the art descriptors for matching, the algorithm performs better than leading feature tracking algorithms including SIFT and SURF in terms of speed and accuracy.

One important parameter of interest in remote sensing to date is yield variability. Proper understanding of yield variability provides insight on the geo-positional dependences of field yields and insight on zone management strategies. Estimating cotton yield and observing cotton yield variability has proven to be a challenging problem due to the complex fruiting behavior of cotton from reactions to environmental conditions. Current methods require expensive sensory equipment on large manned aircrafts and satellites. Other systems, such as cotton yield monitors, are often subject to error due to the collection of dust/trash on photo sensors. This study was aimed towards the development of a miniature unmanned aerial system that utilized a first-person view (FPV) color camera for measuring cotton yield variability. Outcomes of the study led to the development of a method for estimating cotton yield variability from images of experimental cotton plot field taken at harvest time in 2014. These plots were treated with nitrogen fertilizer at five different rates to insure variations in cotton yield across the field. The cotton yield estimates were based on the cotton unit coverage (CUC) observed as the cotton boll image signal density. The cotton boll signals were extracted via their diffusion potential in the image intensity space. This was robust to gradients in illumination caused by cloud coverage as well as fruiting positions in the field. These estimates were provided at a much higher spatial resolution (9.0 cm2) at comparable correlations (R2=0.74) with current expensive systems. This method could prove useful for the development of low cost automated systems for cotton yield estimation as well as yield estimation systems for other crops. / Master of Science

The aim of this project is to design an algorithm that identifies the Area of Interest (AOI) in ELISpot and FluoroSpot images. ELISpot and FluoroSpot are two varieties of a biochemical test used to analyze immune responses by quantifying the amount of cytokine secreted by cells. ELISpot and FluoroSpot images show a well that contains the cytokinesecreting cells which appear as scattered spots. Prior to counting the number of spots, it is required to detect the area in which to count the spots, i.e. the area delimited by the contour of the well. We propose to use the Circle Hough Transform together with filtering and the Laplacian of Gaussian edge detector in order to accurately detect such area. Furthermore we develop an outlier removal method that contributes to increase the robustness of the proposed detection method. Finally we compare our algorithm with another algorithm already in use. A Swedish biotech company called Mabtech has implemented an AOI identifier in the same field. Our proposed algorithm proves to be more robust and provides consistent results for all the images in the dataset.

Computational imaging has been playing an important role in understanding and analysing the captured images. Both image segmentation and restoration has been in-tegral parts of computational imaging. The studies performed in this thesis has been focussed toward developing novel algorithms for image segmentation and restoration. Study related to usage of Morozov Discrepancy Principle in Di use Optical Imaging was also presented here to show that hyper parameter selection could be performed with ease. The Laplacian of Gaussian (LoG) and Canny operators use Gaussian smoothing be-fore applying the derivative operator for edge detection in real images. The LoG kernel was based on second derivative and is highly sensitive to noise when compared to the Canny edge detector. A new edge detection kernel, called as Helmholtz of Gaussian (HoG), which provides higher di suavity is developed in this thesis and it was shown that it is more robust to noise. The formulation of the developed HoG kernel is similar to LoG. It was also shown both theoretically and experimentally that LoG is a special case of HoG. This kernel when used as an edge detector exhibited superior performance compared to LoG, Canny and wavelet based edge detector for the standard test cases both in one- and two-dimensions. The linear inverse problem encountered in restoration of blurred noisy images is typically solved via Tikhonov minimization. The outcome (restored image) of such min-imitation is highly dependent on the choice of regularization parameter. In the absence of prior information about the noise levels in the blurred image, ending this regular-inaction/hyper parameter in an automated way becomes extremely challenging. The available methods like Generalized Cross Validation (GCV) may not yield optimal re-salts in all cases. A novel method that relies on minimal residual method for ending the regularization parameter automatically was proposed here and was systematically compared with the GCV method. It was shown that the proposed method performance was superior to the GCV method in providing high quality restored images in cases where the noise levels are high Di use optical tomography uses near infrared (NIR) light as the probing media to recover the distributions of tissue optical properties with an ability to provide functional information of the tissue under investigation. As NIR light propagation in the tissue is dominated by scattering, the image reconstruction problem (inverse problem) is non-linear and ill-posed, requiring usage of advanced computational methods to compensate this. An automated method for selection of regularization/hyper parameter that incorporates Morozov discrepancy principle(MDP) into the Tikhonov method was proposed and shown to be a promising method for the dynamic Di use Optical Tomography.

In 3D modelling software, deformations are used to add, to remove, or to modify geometric features of existing 3D models to create new models with similar but slightly different details. Traditional techniques for deforming virtual 3D models require users to explicitly define control points and regions of interest (ROIs), and to define precisely how to deform ROIs using control points. The awkwardness of defining these factors in traditional 3D modelling software makes it difficult for people with limited experience of 3D modelling to deform existing 3D models as they expect. As applications which require virtual 3D model processing become more and more widespread, it becomes increasingly desirable to lower the "difficulty of use" threshold of 3D model deformations for users. This thesis argues that the user experience, in terms of intuitiveness and ease of use, of a user interface for deforming virtual 3D models, can be greatly enhanced by employing sketch-based 3D model deformation techniques, which require the minimal quantities of interactions, while keeping the plausibility of the results of deformations as well as the responsiveness of the algorithms, based on modern home grade computing devices. A prototype system for sketch-based 3D model deformations is developed and implemented to support this hypothesis, which allows the user to perform a deformation using a single deforming stroke, eliminating the need to explicitly select control points, the ROI and the deforming operation. GPU based accelerations have been employed to optimise the runtime performance of the system, so that the system is responsive enough for real-time interactions. The studies of the runtime performance and the usability of the prototype system are conducted to provide evidence to support the hypothesis.

The aim of the thesis is detection of a known object in series of pictures. Detection is divided into two methods. First method is based on edge and color detection and comparison. Edge detection is based on detection using both Gradient and Laplacian, so on the first-order and the second-order derivative. Sobel operators were used as well as Laplacian of gaussian method. Thresholding is also used as well as autothreshold calculation. There are two variants of color detection considered in the thesis, direct color comparison and detection based on interest color search. The second part of the thesis is based on interested point detection using a modified SURF method to detect a known object in series of pictures.