This thesis is focused on the design, synthesis and characterization of magnetite-ionic copolymer complexes as nanocarriers for drug delivery and magnetic resonance imaging. The polymers included phosphonate and carboxylate-containing graft and block copolymers. Oleic-acid coated magnetite nanoparticles (8-nm and 16-nm diameters) were investigated. Cisplatin and carboplatin were used as sample drugs. The potentials of the magnetite-ionomer complexes as dual drug delivery carriers and magnetic resonance imaging agents were evaluated. An acrylate-functional poly(ethylene oxide) macromonomer and hexyl (and propyl) ammonium bisphosphonate methacrylate monomers were synthesized. Conventional free radical copolymerizations were conducted to synthesize the graft copolymers. The acrylate-functional poly(ethylene oxide) macromonomer was also used to form graft copolymers with tert-butyl acrylate. Block ionomers containing poly(tert-butyl acrylate) were synthesized via atom transfer radical polymerization, then the tert-butyl groups were removed to afford anions. All the monomers and polymers were characterized by 1H NMR to confirm their structures and assess their compositions. Phosphonate-containing polymers were also characterized by 31P NMR. Magnetite nanoparticles (8-nm diameter) were synthesized by reducing Fe(acac)3 with benzyl alcohol. The 16-nm diameter magnetite was synthesized by thermal decomposition of an iron oleate precursor in trioctylamine as a high-boiling solvent. The iron-oleate precursor was synthesized with iron (III) chloride hexahydrate and sodium oleate with mixed solvents. TEM images of the magnetite were obtained. Magnetite-ionomer complexes were synthesized by binding a portion of the anions (carboxylate or phosphonate) on the copolymers onto the surfaces the magnetite. The remainder of the anions was used to bind with cisplatin and carboplatin via chelation. Physicochemical properties of the complexes were measured by dynamic light scattering. All the complexes with different polymers and magnetite nanoparticles displayed relatively uniform sizes and good size distributions. The magnetite-ionomer complexes displayed good colloidal stabilities in simulated physiological conditions for at least 24 hours. Those graft and block copolymer-magnetite complexes may be good candidates as drug carriers for delivery applications. After cisplatin and carboplatin loading, the sizes of the complexes increased slightly and the zeta potential decreased slightly, which indicated that the loadings were successful. Minimal loss of iron was found, signaling that the binding strengths between the magnetite and the anions of the graft copolymers were strong. 8.7 wt% of platinum was found in the cisplatin loaded complexes and 6.9% in the carboplatin loaded complexes. The results indicated that the magnetite-graft ionomer complexes were capable of loading drugs. Drug release studies were performed at pH 4.6 and 7.4 to mimick endosomal conditions and the physiological environment. Sustained release of drugs was observed. This further indicated the potential for using the magnetite-ionomer complexes as drug carriers. Transverse relaxivities of the magnetite-ionomer complexes with and without drugs were measured and compared to a commercial T2-weighted iron MRI contrast agent-Feridex®. All the complexes had higher relaxivities compared to Feridex®. Thus, the magnetite-ionomer complexes are promising candidates for dual magnetic resonance imaging and drug delivery.Moreover, the aqueous dispersion of the complexes was found to heat upon exposure to an AC magnetic field, thus potentially allowing heat-induced drug release. / Master of Science

Inversion Recovery (IR) is a powerful tool for contrast manipulation in Mag- netic Resonance Imaging (MRI). IR can provide strong contrast between tissues with different values of T1 relaxation times. The tissue magnetization stored at an IR image pixel can take positive as well as negative values. The corresponding polarity information is contained in the phase of the complex image. Due to numerous factors associated with the Magnetic Resonance (MR) scanner and the associated acquisition system, the acquired complex image is modulated by a spatially varying background phase which makes the retrieval of polarity information non-trivial. Many commercial MR scanners perform magnitude-only reconstruction which, due to loss of polarity information, reduces the dynamic contrast range. Phase sensitive IR (PSIR) can provide enhanced image contrast by estimating and removing the background phase and retrieving the correct polarity information. In this thesis, the background phase of complex MR image is modeled using a statistical model based on Markov Ran- dom Fields (MRF). Two model optimization methods have been developed. The first method is a computationally effcient algorithm for finding semi-optimal solutions satisfying the proposed model. Using an adaptive model neighborhood, it can recon- struct low SNR images with slow phase variations. The second method presents a region growing approach which can handle images with rapid phase variations. Ex- perimental results using computer simulations and in vivo experiments show that the proposed method is robust and can perform successful reconstruction even in adverse cases of low signal to noise ratios (SNRs) and high phase variations.

Pedestrian detection is one of the most researched areas in computer vision and is rapidly gaining importance with the emergence of autonomous vehicles and steering assistance technology. Much work has been done in this field, ranging from the collection of extensive datasets to benchmarking of new technologies, but all the research depends on high-quality hardware such as high-resolution cameras, Light Detection and Ranging (LIDAR) and radar. For detection in low-quality moving camera videos, we use image deblurring techniques to reconstruct image frames and use existing pedestrian detection algorithms and compare our results with the leading research done in this area.

Nowadays devices are able to capture and process images from complex surveillance monitoring systems or from simple mobile phones. In certain applications, the time necessary to process the image is not as important as the quality of the processed images (e.g., medical imaging), but in other cases the quality can be sacrificed in favour of time. This thesis focuses on the latter case, and proposes two methodologies for fast image contrast enhancement methods. The proposed methods are based on histogram equalization (HE), and some for handling gray-level images and others for handling color images As far as HE methods for gray-level images are concerned, current methods tend to change the mean brightness of the image to the middle level of the gray-level range. This is not desirable in the case of image contrast enhancement for consumer electronics products, where preserving the input brightness of the image is required to avoid the generation of non-existing artifacts in the output image. To overcome this drawback, Bi-histogram equalization methods for both preserving the brightness and contrast enhancement have been proposed. Although these methods preserve the input brightness on the output image with a significant contrast enhancement, they may produce images which do not look as natural as the ones which have been input. In order to overcome this drawback, we propose a technique called Multi-HE, which consists of decomposing the input image into several sub-images, and then applying the classical HE process to each one of them. This methodology performs a less intensive image contrast enhancement, in a way that the output image presented looks more natural. We propose two discrepancy functions for image decomposition which lead to two new Multi-HE methods. A cost function is also used for automatically deciding in how many sub-images the input image will be decomposed on. Experimental results show that our methods are better in preserving the brightness and producing more natural looking images than the other HE methods. In order to deal with contrast enhancement in color images, we introduce a generic fast hue-preserving histogram equalization method based on the RGB color space, and two instances of the proposed generic method. The first instance uses R-red, G-green, and Bblue 1D histograms to estimate a RGB 3D histogram to be equalized, whereas the second instance uses RG, RB, and GB 2D histograms. Histogram equalization is performed using 7 Abstract 8 shift hue-preserving transformations, avoiding the appearance of unrealistic colors. Our methods have linear time and space complexities with respect to the image dimension, and do not require conversions between color spaces in order to perform image contrast enhancement. Objective assessments comparing our methods and others are performed using a contrast measure and color image quality measures, where the quality is established as a weighed function of the naturalness and colorfulness indexes. This is the first work to evaluate histogram equalization methods with a well-known database of 300 images (one dataset from the University of Berkeley) by using measures such as naturalness and colorfulness. Experimental results show that the value of the image contrast produced by our methods is in average 50% greater than the original image value, and still keeping the quality of the output images close to the original

Enhancing low light videos has been quite a challenge over the years. A video taken in low light always has the issues of low dynamic range and high noise. This master thesis presents contribution within the field of low light video enhancement. Three models are proposed with different tone mapping algorithms for extremely low light low quality video enhancement. For temporal noise removal, a motion compensated kalman structure is presented. Dynamic range of the low light video is stretched using three different methods. In Model 1, dynamic range is increased by adjustment of RGB histograms using gamma correction with a modified version of adaptive clipping thresholds. In Model 2, a shape preserving dynamic range stretch of the RGB histogram is applied using SMQT. In Model 3, contrast enhancement is done using CLAHE. In the final stage, the residual noise is removed using an efficient NLM. The performance of the models are compared on various Objective VQA metrics like NIQE, GCF and SSIM. To evaluate the actual performance of the models subjective tests are conducted, due to the large number of applications that target humans as the end user of the video.The performance of the three models are compared for a total of ten real time input videos taken in extremely low light environment. A total of 25 human observers subjectively evaluated the performance of the three models based on the parameters: contrast, visibility, visually pleasing, amount of noise and overall quality. A detailed statistical evaluation of the relative performance of the three models is also provided.

L'imagerie polarimétrique consiste à acquérir des images contenant des informations relatives à la polarisation de la lumière diffusée par une scène. L'objectif de cette thèse est d'utiliser les propriétés de ce type d'imagerie afin de d'améliorer le contraste entre plusieurs objets d'intérêt.Dans le cadre de l'optimisation du contraste entre deux objets d'intérêt, nous démontrons que, si l'on travaille à temps d'acquisition fixe, c'est l'acquisition d'une unique image avec des états d'illumination et d'analyse optimisés qui permet d'atteindre les meilleures performances. C'est pourquoi nous avons développé un imageur pouvant générer et analyser n'importe quel état de polarisation sur la sphère de Poincaré, en utilisant des cellules à cristaux liquides. Ces états peuvent être contrôlés afin de faire varier le contraste dans les images et nous montrons que les ``états optimaux" permettant de maximiser le contraste dépendent des conditions de mesure. En particulier, la valeur des états de polarisation maximisant le contraste entre deux objets d'intérêt dépend des bruits de mesure (bruit de détecteur, bruit de Poisson, Speckle) ainsi que des fluctuations spatiales des propriétés polarimétriques dans la scène. Une mauvaise estimation de la source de bruit peut donc amener à une perte significative de contraste.Nous nous intéressons ensuite à un scénario d'imagerie plus complexe où la scène peut être illuminée de manière non-uniforme. Nous proposons une méthode d'acquisition utilisant l'ensemble des degrés de liberté fournis par notre imageur et montrons que cette méthode permet d'augmenter significativement le contraste par rapport aux résultats obtenus avec d'autres types d'imagerie comme l'imagerie OSC (Orthogonal State Contrast).Nous étendons ensuite nos études à un cas ``multicibles" où plus de deux objets doivent être distingués. Nous montrons notamment que l'accroissement du nombre d'images peut dégrader le contraste et qu'il existe un nombre optimal d'images à acquérir si l'on travaille à temps d'acquisition fixe.Enfin, nous proposons une méthode visant à automatiser notre imageur pour l'optimisation du contraste en combinant de manière itérative l'acquisition d'images polarimétriques optimisées et un algorithme de segmentation par contours actifs statistiques. Des premiers résultats expérimentaux mettent en évidence l'avantage de cette intégration d'algorithmes de traitement numérique au c\oe ur du processus d'acquisition de l'image. / The polarimetric imaging consists in acquiring images containing information relating to the polarization of the scattered light from a scene. The objective of this thesis is to use the properties of this type of imaging to enhance the contrast between several objects of interest.Considering the optimization of the contrast between two objects of interest, we demonstrate that, if the time for the measurement is fixed, it is the acquisition of a single image with optimized states in illumination and analysis that achieves the best performance. That is why we have developed an imager that can generate and analyze any polarization state on the Poincaré sphere, using liquid crystal cells. These states can be controlled to modify the contrast in the images and we show that the ``optimal states" maximizing the contrast depend on the measurement conditions. Specifically, the value of the polarization states maximizing the contrast between two objects interest depends on the measurement noise (noise detector, Poisson noise, Speckle) and also of spatial fluctuations of polarimetric properties in the scene. Improper estimate of the noise source may therefore lead to a significant loss of contrast.We then consider a more complex imaging scenario where the scene can be illuminated non-uniformly. We propose a method of acquisition using all the degrees of freedom provided by our imaging and show that this method can significantly increase the contrast compared to results obtained with other types of polarimetric imaging such as OSC imaging (Orthogonal State Contrast).We then extend our studies to a ``multi-target case" where more than two objects must be distinguished. In particular, we show that increasing the number of images can degrade the contrast and that there is an optimum number of images to be acquired if one works with a fixed acquisition time.Finally, we propose a method to automate our imaging to optimize contrast by combining iteratively the acquisition of polarimetric images and optimized segmentation algorithm using statistical active contours. The first experimental results demonstrate the advantage of this integration of digital processing algorithms in the core of the image acquisition process.

<p>Xilinx Virtex II Pro FPGA with integrated PowerPC core offers an opportunity to implementing a software and hardware codesign. The software application executes on the PowerPC processor while the FPGA implementation of hardware cores coprocess with PowerPC to achieve the goals of acceleration. Another benefit of coprocessing with the hardware acceleration core is the release of processor load. This thesis demonstrates such an FPGA based software and hardware codesign by implementing a real time video processing project on Xilinx ML310 development platform which is featured with a Xilinx Virtex II Pro FPGA. The software part in this project performs video and memory interface task which includes image capture from camera, the store of image into on-board memory, and the display of image on a screen. The hardware coprocessing core does a contrast enhancement function on the input image. To ease the software development and make this project flexible for future extension, an Embedded Operating System MontaVista Linux is installed on the ML310 platform. Thus the software video interface application is developed using Linux programming method, for example the use of Video4Linux API. The last but not the least implementation topic is the software and hardware interface, which is the Linux device driver for the hardware core. This thesis report presents all the above topics of Operating System installation, video interface software development, contrast enhancement hardware implementation, and hardware core’s Linux device driver programming. After this, a measurement result is presented to show the performance of hardware acceleration and processor load reduction, by comparing to the results from a software implementation of the same contrast enhancement function. This is followed by a discussion chapter, including the performance analysis, current design’s limitations and proposals for improvements. This report is ended with an outlook from this master thesis.</p>

Based on the acquisition trajectory, PROPELLER MRI repeatedly sampled the center k-space in every blade, which was used to provide most of the energy of an image. The purpose of view sharing PROPELLER is to improve the spatio-temporal resolution of dynamic imaging by reducing the acquisition time of single frame to that of single blade. With the center k-space provided by only one blade, which is called the target blade, the high spatial-frequency components were appropriately contributed by a set of neighboring blades with different rotation angles, leading to the high spatial resolution after reconstruction. In this study, a flow phantom experiment with the injection of T1-shortening Gd-DTPA solution was performed to exam the feasibility and accuracy of view-sharing PROPELLER. Furthermore, cardiac imaging of healthy volunteer obtained by the proposed technique was also done with ECG gating to test the image quality without any injection of contrast agent. The in-vivo experiment was done with and without breath holding. In addition to slight aliasing artifact due to insufficient FOV, no other artifact was observed.

Today, infrared cameras are used especially for target tracking and surveillance operations. However, they have a high dynamic range output, and the standard display devices cannot handle them. In order to show them on common devices, the dynamic range is cropped. Thus, the contrast of the image is reduced. This is called as the High Dynamic Range (HDR) Compression. Although several algorithms have been proposed for preserving details during the HDR compression process, it cannot be used to enhance the local contrasts of image contents. In this thesis, we compare the performances of contrast enhancement techniques, which are suitable for real time applications. The methods experimented are generally histogram based methods. Some modifications are also proposed in order to reduce computational complexity of the process. Performances of these methods are compared with common objective quality metrics on different image sets.

Xilinx Virtex II Pro FPGA with integrated PowerPC core offers an opportunity to implementing a software and hardware codesign. The software application executes on the PowerPC processor while the FPGA implementation of hardware cores coprocess with PowerPC to achieve the goals of acceleration. Another benefit of coprocessing with the hardware acceleration core is the release of processor load. This thesis demonstrates such an FPGA based software and hardware codesign by implementing a real time video processing project on Xilinx ML310 development platform which is featured with a Xilinx Virtex II Pro FPGA. The software part in this project performs video and memory interface task which includes image capture from camera, the store of image into on-board memory, and the display of image on a screen. The hardware coprocessing core does a contrast enhancement function on the input image. To ease the software development and make this project flexible for future extension, an Embedded Operating System MontaVista Linux is installed on the ML310 platform. Thus the software video interface application is developed using Linux programming method, for example the use of Video4Linux API. The last but not the least implementation topic is the software and hardware interface, which is the Linux device driver for the hardware core. This thesis report presents all the above topics of Operating System installation, video interface software development, contrast enhancement hardware implementation, and hardware core’s Linux device driver programming. After this, a measurement result is presented to show the performance of hardware acceleration and processor load reduction, by comparing to the results from a software implementation of the same contrast enhancement function. This is followed by a discussion chapter, including the performance analysis, current design’s limitations and proposals for improvements. This report is ended with an outlook from this master thesis.