Matching Items (5)
Filtering by
- All Subjects: Computer vision
- Creators: Frakes, David
Description
Image understanding has been playing an increasingly crucial role in vision applications. Sparse models form an important component in image understanding, since the statistics of natural images reveal the presence of sparse structure. Sparse methods lead to parsimonious models, in addition to being efficient for large scale learning. In sparse modeling, data is represented as a sparse linear combination of atoms from a "dictionary" matrix. This dissertation focuses on understanding different aspects of sparse learning, thereby enhancing the use of sparse methods by incorporating tools from machine learning. With the growing need to adapt models for large scale data, it is important to design dictionaries that can model the entire data space and not just the samples considered. By exploiting the relation of dictionary learning to 1-D subspace clustering, a multilevel dictionary learning algorithm is developed, and it is shown to outperform conventional sparse models in compressed recovery, and image denoising. Theoretical aspects of learning such as algorithmic stability and generalization are considered, and ensemble learning is incorporated for effective large scale learning. In addition to building strategies for efficiently implementing 1-D subspace clustering, a discriminative clustering approach is designed to estimate the unknown mixing process in blind source separation. By exploiting the non-linear relation between the image descriptors, and allowing the use of multiple features, sparse methods can be made more effective in recognition problems. The idea of multiple kernel sparse representations is developed, and algorithms for learning dictionaries in the feature space are presented. Using object recognition experiments on standard datasets it is shown that the proposed approaches outperform other sparse coding-based recognition frameworks. Furthermore, a segmentation technique based on multiple kernel sparse representations is developed, and successfully applied for automated brain tumor identification. Using sparse codes to define the relation between data samples can lead to a more robust graph embedding for unsupervised clustering. By performing discriminative embedding using sparse coding-based graphs, an algorithm for measuring the glomerular number in kidney MRI images is developed. Finally, approaches to build dictionaries for local sparse coding of image descriptors are presented, and applied to object recognition and image retrieval.
ContributorsJayaraman Thiagarajan, Jayaraman (Author) / Spanias, Andreas (Thesis advisor) / Frakes, David (Committee member) / Tepedelenlioğlu, Cihan (Committee member) / Turaga, Pavan (Committee member) / Arizona State University (Publisher)
Created2013
Description
There is a growing interest for improved high-accuracy camera calibration methods due to the increasing demand for 3D visual media in commercial markets. Camera calibration is used widely in the fields of computer vision, robotics and 3D reconstruction. Camera calibration is the first step for extracting 3D data from a 2D image. It plays a crucial role in computer vision and 3D reconstruction due to the fact that the accuracy of the reconstruction and 3D coordinate determination relies on the accuracy of the camera calibration to a great extent. This thesis presents a novel camera calibration method using a circular calibration pattern. The disadvantages and issues with existing state-of-the-art methods are discussed and are overcome in this work. The implemented system consists of techniques of local adaptive segmentation, ellipse fitting, projection and optimization. Simulation results are presented to illustrate the performance of the proposed scheme. These results show that the proposed method reduces the error as compared to the state-of-the-art for high-resolution images, and that the proposed scheme is more robust to blur in the imaged calibration pattern.
ContributorsPrakash, Charan Dudda (Author) / Karam, Lina J (Thesis advisor) / Frakes, David (Committee member) / Papandreou-Suppappola, Antonia (Committee member) / Arizona State University (Publisher)
Created2012
Description
Diabetic retinopathy (DR) is a common cause of blindness occurring due to prolonged presence of diabetes. The risk of developing DR or having the disease progress is increasing over time. Despite advances in diabetes care over the years, DR remains a vision-threatening complication and one of the leading causes of blindness among American adults. Recent studies have shown that diagnosis based on digital retinal imaging has potential benefits over traditional face-to-face evaluation. Yet there is a dearth of computer-based systems that can match the level of performance achieved by ophthalmologists. This thesis takes a fresh perspective in developing a computer-based system aimed at improving diagnosis of DR images. These images are categorized into three classes according to their severity level. The proposed approach explores effective methods to classify new images and retrieve clinically-relevant images from a database with prior diagnosis information associated with them. Retrieval provides a novel way to utilize the vast knowledge in the archives of previously-diagnosed DR images and thereby improve a clinician's performance while classification can safely reduce the burden on DR screening programs and possibly achieve higher detection accuracy than human experts. To solve the three-class retrieval and classification problem, the approach uses a multi-class multiple-instance medical image retrieval framework that makes use of spectrally tuned color correlogram and steerable Gaussian filter response features. The results show better retrieval and classification performances than prior-art methods and are also observed to be of clinical and visual relevance.
ContributorsChandakkar, Parag Shridhar (Author) / Li, Baoxin (Thesis advisor) / Turaga, Pavan (Committee member) / Frakes, David (Committee member) / Arizona State University (Publisher)
Created2012
Description
This paper summarizes the [1] ideas behind, [2] needs, [3] development, and [4] testing of 3D-printed sensor-stents known as Stentzors. This sensor was successfully developed entirely from scratch, tested, and was found to have an output of 3.2*10-6 volts per RMS pressure in pascals. This paper also recommends further work to render the Stentzor deployable in live subjects, including [1] further design optimization, [2] electrical isolation, [3] wireless data transmission, and [4] testing for aneurysm prevention.
ContributorsMeidinger, Aaron Michael (Author) / LaBelle, Jeffrey (Thesis director) / Frakes, David (Committee member) / Barrett, The Honors College (Contributor) / Mechanical and Aerospace Engineering Program (Contributor)
Created2014-05
Description
Intracranial aneurysms, which form in the blood vessels of the brain, are particularly dangerous because of the importance and fragility of the human brain. When an intracranial aneurysm gets large it poses a significant risk of bursting and causing subarachnoid hemorrhaging (SAH), a possibly fatal condition. One possible treatment involves placing a stent in the vessel to act as a flow diverter. In this study we look at the hemodynamics of two geometries of idealized basilar tip aneurysms, at 2,3, and 4 ml/s pulsatile flow, at three different points in the cardiac cycle. The smaller model had neck and dome diameters of 2.67 mm and 4 mm respectively, while the larger aneurysm had neck and dome diameters of 3 mm and 6 mm respectively. Both diameters and the dome to neck ratio increased in the second model, representing growth over time. Flow was analyzed using stereoscopic particle image velocimetry (PIV) for both geometries in untreated models, as well as after treatment with a high porosity Enterprise stent (Codman and Shurtleff Inc.). Flow in the models was characterized by root mean square velocity in the aneurysm and neck plane, cross neck flow, max aneurysm vorticity, and total aneurysm kinetic energy. It was found that in the smaller aneurysm model (model 1), Enterprise stent treatment reduced all flow parameters substantially. The smallest reduction was in max vorticity, at 42.48%, and the largest in total kinetic energy, at 75.69%. In the larger model (model 2) there was a 52.18% reduction in cross neck flow, but a 167.28% increase in aneurysm vorticity. The other three parameters experienced little change. These results, along with observed velocity vector fields, indicate a noticeable diversion of flow away from the aneurysm in the stent treated model 1. Treatment in model 2 had a small flow diversion effect, but also altered flow in unpredictable ways, in some cases having a detrimental effect on aneurysm hemodynamics. The results of this study indicate that Enterprise stent treatment is only effective in small, relatively undeveloped aneurysm geometries, and waiting until an aneurysm has grown too large can eliminate this treatment option altogether.
ContributorsLindsay, James Bryan (Author) / Frakes, David (Thesis director) / LaBelle, Jeffrey (Committee member) / Nair, Priya (Committee member) / Barrett, The Honors College (Contributor) / School of Humanities, Arts, and Cultural Studies (Contributor)
Created2013-05