Dissertation (PhD)--University of Stellenbosch, 2001. / ENGLISH ABSTRACT: The purpose of this investigation was to explore the multiple variables involved in TACT® image generation in an attempt to optimize this imaging modality for the diagnostic task of primary dental caries detection. The work is divided in seven phases in which the variables are evaluated individually. Teeth from the study samples were mounted in dental stone and imaged with a solid state digital radiography sensor. As a requisite of TACT® imaging, multiple images of the teeth were acquired from different projection angles. These resulting basis images were then used to generate TACT® slices. Variables tested in the investigation included the number of iterative restorations to which the slices were submitted, the number of basis images, the angle formed between the basis images, the two- and three-dimensional distribution of the basis projections in space, and the method through which the slices were reconstructed. For all phases, observers were asked to assess the presence or absence of primary caries in the teeth imaged using the TACT® slices treated with the different variables. Finally, to determine whether the best combination of variables produced a significant improvement in diagnostic performance, a comparison with conventional digital radiography images was carried out. No statistically significant differences were found in caries detection between TACT® slices submitted to different numbers of iterative restorations, reconstructed from basis images bearing different angular disparities, spatial distributions (in both two and three dimensions), or through different reconstruction methods. A statistically significant difference was detected between TACT® slices reconstructed from different numbers of basis projections. The final comparison showed that TACT® was not statistically superior to conventional digital radiography for the task of caries detection. The results of this investigation suggest that, although TACT® has been shown to be useful in many tasks performed in dentistry, its application in caries detection is not essential inasmuch as there are modalities that are simpler, more practical, less expensive, and that submit the patient to smaller radiation doses. Keywords: TACT, tomosynthesis, image reconstruction, digital radiography, caries detection, ROC analysis, analysis of variance / AFRIKAANSE OPSOMMING: Die doel van hierdie proefskrif was om die veelvuldige veranderlikes wat betrokke is by gewysigde spleet rekenaartomografie (Tuned Aperture Computed Tomography (TACT)) te ondersoek in 'n poging om die beeldingsmodaliteit te optimaliseer in die diagnostiese opsporing van primere karies. Die proefskrif bestaan uit 'n ekstensiewe literatuur oorsig en word in 7 fases aangebied waarin die veranderlikes individueel geevalueer word. Tande is in gips ingebed en radiografiese opnames is gemaak met behulp van 'n digitale radiografiese sensor. As 'n voorvereiste vir TACT beelding is veelvuldige beelde uit verskillende projeksiehoeke van die tande gemaak. Die resulterende basisbeelde is dan gebruik om TACT snitte te produseer. Veranderlikes wat in die proefskrif getoets is, sluit die volgende in: 'n aantal herhalende herstellings waaraan die snitte blootgestel is, die aantal basisbeelde, die hoek gevorm tussen die basisbeelde, die 2 en 3 dimensionele verspeiding van die basis projeksies in die ruimte en die metodes waardeur die snitte gerekonstrueer is. In aIle fases is waarnemers gevra om die teenwoordigheid of afwesigheid van primere karies te evalueer wat met TACT afgeneem is met in ag neming van die verskillende veranderlikes. Ten slotte, om te bepaal of die beste kombinasie van veranderlikes 'n aansienlike verbetering in diagnostiese prestasies sou meebring, is 'n vergelyking met konvensionele digitale radiografiese beelding uitgevoer. Geen statistiese beduidende verskille is waargeneem in die opsporing van karies tussen TACT snitte wat blootgestel is aan verskillende aantal herhalende herstellings, rekonstruksie van basis beelding met verskillende hoek veranderinge, ruimtelike verspreiding (beide 2 en 3 dimensioneel) of deur verskillende rekonstruksie metodes nie. 'n Statistiese beduidende verskil is waargeneem tussen TACT snitte wat van 'n verskeidenheid basis projeksies gerekonstrueer is. Die finale vergelyking het aangetoon dat TACT nie statisties beter is as konvensionele radiografie in die opsporing van karies nie. Die resultate van hierdie proefskrif het getoon dat alhoewel TACT bruikbaar is in vele prosedures wat in die tandheelkunde uitgevoer word, is die toepassing daarvan in die diagnose van karies nie noodsaaklik nie, omdat daar in die tandheelkunde modaliteite beskikbaar is, wat meer eenvoudig, meer prakties en goedkoper is, met 'n laer stralingsdosis vir die pasient.

Thesis (PhD (Electronic Engineering))--University of Stellenbosch, 2009. / The downloading of images is a bottleneck in the image acquisition chain for low earth orbit, remote sensing satellites. An on-board image quality assessment system could optimise use of available downlink time by prioritising images for download, based on their quality. An image quality assessment system based on measuring image degradations is proposed. Algorithms for estimating degradations are investigated. The degradation types considered are cloud cover, additive sensor noise and the defocus extent of the telescope. For cloud detection, the novel application of heteroscedastic discriminant analysis resulted in better performance than comparable dimension reducing transforms from remote sensing literature. A region growing method, which was previously used on-board a micro-satellite for cloud cover estimation, is critically evaluated and compared to commonly used thresholding. The thresholding method is recommended. A remote sensing noise estimation algorithm is compared to a noise estimation algorithm based on image pyramids. The image pyramid algorithm is recommended. It is adapted, which results in smaller errors. A novel angular spectral smoothing method for increasing the robustness of spectral based, direct defocus estimation is introduced. Three existing spectral based defocus estimation methods are compared with the angular smoothing method. An image quality assessment model is developed that models the mapping of the three estimated degradation levels to one quality score. A subjective image quality evaluation experiment is conducted, during which more than 18000 independent human judgements are collected. Two quality assessment models, based on neural networks and splines, are tted to this data. The spline model is recommended. The integrated system is evaluated and image quality predictions are shown to correlate well with human quality perception.

With the advance of the LCD technology, video quality is becoming increasingly important. In this thesis, we develop hardware-friendly low-complexity enhancement algorithms. Video quality enhancement methods can be classified into two main categories. Single frame methods are the first category. These methods have generally low computational complexity. Multi-frame methods combine information from more than one frame and require the motion information of objects in the scene to do so. We first concentrate on the contrast-enhancement problem by using both global (frame-wise) and local information derived from the image. We use the image histogram and present a regularization-based histogram modification method to avoid problems that are often created by histogram equalization. Next, we design a compression artifact reduction algorithm that reduces ringing artifacts, which is disturbing especially on large displays. Furthermore, to remove the blurriness in the original video we present a non-iterative diffusion-based sharpening algorithm, which enhances edges in a ringing-aware fashion. The diffusion-based technique works on gradient approximations in a neighborhood individually. This gives more freedom compared to modulating the high-pass filter output that is used to sharpen the edges. Motion estimation enables applications such as motion-compensated noise reduction, frame-rate conversion, de-interlacing, compression, and super-resolution. Motion estimation is an ill-posed problem and therefore requires the use of prior knowledge on motion of objects. Objects have inertia and are usually larger then pixels or a block of pixels in size, which creates spatio-temporal correlation. We design a method that uses temporal redundancy to improve motion-vector search by choosing bias vectors from the previous frame and adaptively penalizes deviations from the bias vectors. This increases the robustness of the motion-vector search. The spatial correlation is more reliable because temporal correlation is difficult to use when the objects move fast or accelerate in time, or have small sizes. Spatial smoothness is not valid across motion boundaries. We investigate using energy minimization for motion estimation and incorporate the spatial smoothness prior into the energy. By formulating the energy minimization iterations for each motion vector as the primal problem, we show that the dual problem is motion segmentation for that specific motion vector.

This dissertation develops several new algorithms to solve existing problems in practical application of the previously developed PCA+JPEG2000, which has shown superior rate-distortion performance in hyperspectral image compression. In addition, a new scheme is proposed to facilitate multi-temporal hyperspectral image compression. Specifically, the uniqueness in each algorithm is described as follows. 1. An empirical piecewise linear equation is proposed to estimate the optimal number of major principal components (PCs) used in SubPCA+JPEG2000 for AVIRIS data. Sensor-specific equations are presented with excellent fitting performance for AVIRIS, HYDICE, and HyMap data. As a conclusion, a general guideline is provided for finding sensor-specific piecewise linear equations. 2. An anomaly-removal-based hyperspectral image compression algorithm is proposed. It preserves anomalous pixels in a lossless manner, and yields the same or even improved rate-distortion performance. It is particularly useful to SubPCA+JPEG2000 when compressing data with anomalies that may reside in minor PCs. 3. A segmented PCA-based PCA+JPEG2000 compression algorithm is developed, which spectrally partitions an image based on its spectral correlation coefficients. This compression scheme greatly improves the rate-distortion performance of PCA+JPEG2000 when the spatial size of the data is relatively smaller than its spectral size, especially at low bitrates. A sensor-specific partition method is also developed for fast processing with suboptimal performance. 4. A joint multi-temporal image compression scheme is proposed. The algorithm preserves change information in a lossless fashion during the compression. It can yield perfect change detection with slightly degraded rate-distortion performance.

The purpose of this project is to develop a Java Applet that provides all the tools needed for creating image fantasies. It lets the user pick a template and an image, and combine them together. The user can then apply image processing techniques such as rotation, zooming, blurring etc according to his/her requirements.

A avaliação de qualidade de vídeo possui um papel fundamental no processamento de vídeo e em aplicações de comunicação. Uma métrica de qualidade de vídeo ideal deve garantir a alta correlação entre a predição da distorção do vídeo e a percepção de qualidade do Sistema Visual Humano. Este trabalho propõe o uso de modelos de atenção visual com abordagem bottom up baseados em saliências para avaliação de qualidade de vídeo. Três métricas objetivas de avaliação são propostas. O primeiro método é uma métrica com referência completa baseada na estrutura de similaridade. O segundo modelo é uma métrica sem referência baseada em uma modelagem sigmoidal com solução de mínimos quadrados que usa o algoritmo de Levenberg-Marquardt e extração de características espaço-temporais. E, a terceira métrica é análoga à segunda, porém usa a característica Blockiness na detecção de distorções de blocagem no vídeo. A abordagem bottom-up é utilizada para obter os mapas de saliências que são extraídos através de um modelo multiescala de background baseado na detecção de movimentos. Os resultados experimentais apresentam um aumento da eficiência de predição de qualidade de vídeo nas métricas que utilizam o modelo de saliência em comparação com as respectivas métricas que não usam este modelo, com destaque para as métricas sem referência propostas que apresentaram resultados melhores do que métricas com referência para algumas categorias de vídeos. / Video quality assessment plays a key role in the video processing and communications applications. An ideal video quality metric shall ensure high correlation between the video distortion prediction and the perception of the Human Visual System. This work proposes the use of visual attention models with bottom-up approach based on saliencies for video qualitty assessment. Three objective metrics are proposed. The first method is a full reference metric based on the structural similarity. The second is a no reference metric based on a sigmoidal model with least squares solution using the Levenberg-Marquardt algorithm and extraction of spatial and temporal features. And, the third is analagous to the last one, but uses the characteristic Blockiness for detecting blocking distortions in the video. The bottom-up approach is used to obtain the salient maps, which are extracted using a multiscale background model based on motion detection. The experimental results show an increase of efficiency in the quality prediction of the proposed metrics using salient model in comparission to the same metrics not using these model, highlighting the no reference proposed metrics that had better results than metrics with reference to some categories of videos.

A avaliação de qualidade de vídeo possui um papel fundamental no processamento de vídeo e em aplicações de comunicação. Uma métrica de qualidade de vídeo ideal deve garantir a alta correlação entre a predição da distorção do vídeo e a percepção de qualidade do Sistema Visual Humano. Este trabalho propõe o uso de modelos de atenção visual com abordagem bottom up baseados em saliências para avaliação de qualidade de vídeo. Três métricas objetivas de avaliação são propostas. O primeiro método é uma métrica com referência completa baseada na estrutura de similaridade. O segundo modelo é uma métrica sem referência baseada em uma modelagem sigmoidal com solução de mínimos quadrados que usa o algoritmo de Levenberg-Marquardt e extração de características espaço-temporais. E, a terceira métrica é análoga à segunda, porém usa a característica Blockiness na detecção de distorções de blocagem no vídeo. A abordagem bottom-up é utilizada para obter os mapas de saliências que são extraídos através de um modelo multiescala de background baseado na detecção de movimentos. Os resultados experimentais apresentam um aumento da eficiência de predição de qualidade de vídeo nas métricas que utilizam o modelo de saliência em comparação com as respectivas métricas que não usam este modelo, com destaque para as métricas sem referência propostas que apresentaram resultados melhores do que métricas com referência para algumas categorias de vídeos. / Video quality assessment plays a key role in the video processing and communications applications. An ideal video quality metric shall ensure high correlation between the video distortion prediction and the perception of the Human Visual System. This work proposes the use of visual attention models with bottom-up approach based on saliencies for video qualitty assessment. Three objective metrics are proposed. The first method is a full reference metric based on the structural similarity. The second is a no reference metric based on a sigmoidal model with least squares solution using the Levenberg-Marquardt algorithm and extraction of spatial and temporal features. And, the third is analagous to the last one, but uses the characteristic Blockiness for detecting blocking distortions in the video. The bottom-up approach is used to obtain the salient maps, which are extracted using a multiscale background model based on motion detection. The experimental results show an increase of efficiency in the quality prediction of the proposed metrics using salient model in comparission to the same metrics not using these model, highlighting the no reference proposed metrics that had better results than metrics with reference to some categories of videos.

Indiana University-Purdue University Indianapolis (IUPUI) / In crime scene investigations it is necessary to capture images of impression evidence such as tire track or shoe impressions. Currently, such evidence is captured by taking two-dimensional (2D) color photographs or making a physical cast of the impression in order to capture the three-dimensional (3D) structure of the information. This project aims to build a digitizing device that scans the impression evidence and generates (i) a high resolution three-dimensional (3D) surface image, and (ii) a co-registered two-dimensional (2D) color image. The method is based on active structured lighting methods in order to extract 3D shape information of a surface. A prototype device was built that uses an assembly of two line laser lights and a high-definition video camera that is moved at a precisely controlled and constant speed along a mechanical actuator rail in order to scan the evidence. A prototype software was also developed which implements the image processing, calibration, and surface depth calculations. The methods developed in this project for extracting the digitized 3D surface shape and 2D color images include (i) a self-contained calibration method that eliminates the need for pre-calibration of the device; (ii) the use of two colored line laser lights projected from two different angles to eliminate problems due to occlusions; and (iii) the extraction of high resolution color image of the impression evidence with minimal distortion.The system results in sub-millimeter accuracy in the depth image and a high resolution color image that is registered with the depth image. The system is particularly suitable for high quality images of long tire track impressions without the need for stitching multiple images.