Electrical & Computer Engineering Theses and Dissertations
Permanent URI for this collectionhttp://hdl.handle.net/1903/2765
Browse
4 results
Search Results
Item Anomaly Detection in Noisy Images(2015) Gibert Serra, Xavier; Chellappa, Rama; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Finding rare events in multidimensional data is an important detection problem that has applications in many fields, such as risk estimation in insurance industry, finance, flood prediction, medical diagnosis, quality assurance, security, or safety in transportation. The occurrence of such anomalies is so infrequent that there is usually not enough training data to learn an accurate statistical model of the anomaly class. In some cases, such events may have never been observed, so the only information that is available is a set of normal samples and an assumed pairwise similarity function. Such metric may only be known up to a certain number of unspecified parameters, which would either need to be learned from training data, or fixed by a domain expert. Sometimes, the anomalous condition may be formulated algebraically, such as a measure exceeding a predefined threshold, but nuisance variables may complicate the estimation of such a measure. Change detection methods used in time series analysis are not easily extendable to the multidimensional case, where discontinuities are not localized to a single point. On the other hand, in higher dimensions, data exhibits more complex interdependencies, and there is redundancy that could be exploited to adaptively model the normal data. In the first part of this dissertation, we review the theoretical framework for anomaly detection in images and previous anomaly detection work done in the context of crack detection and detection of anomalous components in railway tracks. In the second part, we propose new anomaly detection algorithms. The fact that curvilinear discontinuities in images are sparse with respect to the frame of shearlets, allows us to pose this anomaly detection problem as basis pursuit optimization. Therefore, we pose the problem of detecting curvilinear anomalies in noisy textured images as a blind source separation problem under sparsity constraints, and propose an iterative shrinkage algorithm to solve it. Taking advantage of the parallel nature of this algorithm, we describe how this method can be accelerated using graphical processing units (GPU). Then, we propose a new method for finding defective components on railway tracks using cameras mounted on a train. We describe how to extract features and use a combination of classifiers to solve this problem. Then, we scale anomaly detection to bigger datasets with complex interdependencies. We show that the anomaly detection problem naturally fits in the multitask learning framework. The first task consists of learning a compact representation of the good samples, while the second task consists of learning the anomaly detector. Using deep convolutional neural networks, we show that it is possible to train a deep model with a limited number of anomalous examples. In sequential detection problems, the presence of time-variant nuisance parameters affect the detection performance. In the last part of this dissertation, we present a method for adaptively estimating the threshold of sequential detectors using Extreme Value Theory on a Bayesian framework. Finally, conclusions on the results obtained are provided, followed by a discussion of possible future work.Item Edge-Based Automated Facial Blemish Removal(2013) NessAiver, Avisha; Chellappa, Rama; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This thesis presents an end-to-end approach for taking a an image of a face and seamlessly isolating and filling in any blemishes contained therein. This consists of detecting the face within a larger image, building an accurate mask of the facial features so as not to mistake them as blemishes, detecting the blemishes themselves and painting over them with accurate skin tones. We devote the first part of the thesis to detailing our algorithm for extracting facial features. This is done by first improving the image through histogram equal- ization and illumination compensation followed by finding the features themselves from a computed edge map. Geometric knowledge of general feature positioning and blemish shapes is used to determine which edge clusters belong to correspond- ing facial features. Color and reflectance thresholding is then used to build a skin map. In the second part of the thesis we identify the blemishes themselves. A Lapla- cian of Gaussian blob detector is used to identify potential candidates. Thresholding and dilating operations are then performed to trim this candidate list down followed by the use of various morphological properties to reject regions likely to not be blem- ishes. Finally, in the third part, we examine four possible techniques for inpainting blemish regions once found. We settle on using a technique that fills in pixels based on finding a patch in the nearby image region with the most similar surrounding texture to the target pixel. Priority in the pixel fill-order is given to strong edges and contours.Item Cellular Pattern Quantication and Automatic Bench-marking Data-set Generation on confocal microscopy images(2010) Cui, Chi; JaJa, Joseph; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)The distribution, directionality and motility of the actin fibers control cell shape, affect cell function and are different in cancer versus normal cells. Quantification of actin structural changes is important for further understanding differences between cell types and for elucidation the effects and dynamics of drug interactions. We propose an image analysis framework to quantify the F-actin organization patterns in response to different pharmaceutical treatments.The main problems addressed include which features to quantify and what quantification measurements to compute when dealing with unlabeled confocal microscopy images. The resultant numerical features are very effective to profile the functional mechanism and facilitate the comparison of different drugs. The analysis software is originally implemented in Matlab and more recently the most time consuming part in the feature extraction stage is implemented onto the NVIDIA GPU using CUDA where we obtain 15 to 20 speedups for different sizes of image. We also propose a computational framework for generating synthetic images for validation purposes. The validation for the feature extraction is done by visual inspection and the validation for quantification is done by comparing them with well-known biological facts. Future studies will further validate the algorithms, and elucidate the molecular pathways and kinetics underlying the F-actin changes. This is the first study quantifying different structural formations of the same protein in intact cells. Since many anti-cancer drugs target the cytoskeleton, we believe that the quantitative image analysis method reported here will have broad applications to understanding the mechanisms of candidate pharmaceutical.Item SIMULTANEOUS MULTI-VIEW FACE TRACKING AND RECOGNITION IN VIDEO USING PARTICLE FILTERING(2009) Seo, Naotoshi; Chellappa, Rama; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Recently, face recognition based on video has gained wide interest especially due to its role in surveillance systems. Video-based recognition has superior advantages over image-based recognition because a video contains image sequences as well as temporal information. However, surveillance videos are generally of low-resolution and contain faces mostly in non-frontal poses. We propose a multi-view, video-based face recognition algorithm using the Bayesian inference framework. This method represents an appearance of each subject by a complex nonlinear appearance manifold expressed as a collection of simpler pose manifolds and the connections, represented by transition probabilities, among them. A Bayesian inference formulation is introduced to utilize the temporal information in the video via the transition probabilities among pose manifolds. The Bayesian inference formulation realizes video-based face recognition by progressively accumulating the recognition confidences in frames. The accumulation step possibly enables to solve face recognition problems in low-resolution videos, and the progressive characteristic is especially useful for a real-time processing. Furthermore, this face recognition framework has another characteristic that does not require processing all frames in a video if enough recognition confidence is accumulated in an intermediate frame. This characteristic gives an advantage over batch methods in terms of a computational efficiency. Furthermore, we propose a simultaneous multi-view face tracking and recognition algorithm. Conventionally, face recognition in a video is performed in tracking-then-recognition scenario that extracts the best facial image patch in the tracking and then recognizes the identity of the facial image. Simultaneous face tracking and recognition works in a different fashion, by handling both tracking and recognition simultaneously. Particle filter is a technique for implementing a Bayesian inference filter by Monte Carlo simulation, which has gained prevalence in the visual tracking literature since the Condensation algorithm was introduced. Since we have proposed a video-based face recognition algorithm based on the Bayesian inference framework, it is easy to integrate the particle filter tracker and our proposed recognition method into one, using the particle filter for both tracking and recognition simultaneously. This simultaneous framework utilizes the temporal information in a video for not only tracking but also recognition by modeling the dynamics of facial poses. Although the time series formulation remains more general, only the facial pose dynamics is utilized for recognition in this thesis.