Information retrieval systems are built to handle texts as topical items: texts are tabulated by occurrence frequencies of content words in them, under the assumption that text topic is reasonably well modeled by content word occurrence. But texts have several interesting characteristics beyond topic. The experiments described in this text investigate stylistic variation. Roughly put, style is the difference between two ways of saying the same thing -- and systematic stylistic variation can be used to characterize the genre of documents. These experiments investigate if stylistic information is distinguishable using simple language engineering methods, and if in that case this type of information can be used to improve information retrieval systems.

Abstract—We present a generic detection/localization algorithm capable of searching for a visual object of interest without training. The proposed method operates using a single example of an object of interest to find similar matches; does not require prior knowledge (learning) about objects being sought; and does not require any pre-processing step or segmentation of a target image. Our method is based on the computation of local regression kernels as descriptors from a query, which measure the likeness of a pixel to its surroundings. Salient features are extracted from said descriptors and compared against analogous features from the target image. This comparison is done using a matrix generalization of the cosine similarity measure. We illustrate optimality properties of the algorithm using a naive-Bayes framework. The algorithm yields a scalar resemblance map, indicating the likelihood of similarity between the query and all patches in the target image. By employing nonparametric significance tests and non-maxima suppression, we detect the presence and location of objects similar to the given query. The approach is extended to account for large variations in scale and rotation. High performance is demonstrated on several challenging datasets, indicating successful detection of objects in diverse contexts and under different imaging conditions. Index Terms—Object detection, image representation, correlation and regression analysis 1

We present an algorithm for detecting human actions based upon a single given video example of such actions. The proposed method is unsupervised, does not require learning, segmentation, or motion estimation. The novel features employed in our method are based on space-time locally adaptive regression kernels. Our method is based on the dense computation of so-called space-time local regression kernels (i.e. local descriptors) from a query video, which measure the likeness of a voxel to its spatiotemporal surroundings. Salient features are then extracted from these descriptors using principal components analysis (PCA). These are efficiently compared against analogous features from the target video using a matrix generalization of the cosine similarity measure. The algorithm yields a scalar resemblance volume; each voxel indicating the likelihood of similarity between the query video and all cubes in the target video. By employing non-parametric significance tests and non-maxima suppression, we accurately detect the presence and location of actions similar to the given query video. High performance is demonstrated on a challenging set of action data [8] indicating successful detection of multiple complex actions even in the presence of fast motions. 1.

We present a novel approach to the problem of detection of visual similarity between a template image, and patches in a given image. The method is based on the computation of a local kernel from the template, which measures the likeness of a pixel to its surroundings. This kernel is then used as a descriptor from which features are extracted and compared against analogous features from the target image. Comparison of the features extracted is carried out using canonical correlations analysis. The overall algorithm yields a scalar resemblance map (RM) which indicates the statistical likelihood of similarity between a given template and all target patches in an image being examined. Performing a statistical test on the resulting RM identifies similar objects with high accuracy and is robust to various challenging conditions such as partial occlusion, and illumination change. Index Terms — object detection, local metric learning, kernel regression, canonical correlation analysis, test statistic, principal component analysis

Implementation of the Toxic Substances Control Act of 1977 creates the need to reliably establish testing priorities because laboratory resources are limited and the number of industrial chemicals requiring evaluation is overwhelming. The use of quantitative structure activity relationship (QSAR) models as rapid and predictive screening tools to select more potentially hazardous chemicals for in-depth laboratory evaluation has been proposed. Further implementation and refinement of quantitative structure-toxicity relationships in aquatic toxicology and hazard assessment requires the development of a "mode-of-action" database. With such a database, a qualitative structure-activity relationship can be formulated to assign the proper mode of action, and respective QSAR, to a given chemical structure. In this review, the development of fish acute toxicity syndromes (FATS), which are toxic-response sets based on various behavioral and physiological-biochemical measurements, and their projected use in the mode-of-action database are outlined. Using behavioral parameters monitored in the fathead minnow during acute toxicity testing, FATS associated with acetylcholinesterase (AChE) inhibitors and narcotics could be reliably predicted. However, compounds classified as oxidative phosphorylation uncouplers or stimulants could not be resolved. Refinement of this approach by using respiratory-cardiovascular responses in the rainbow trout, enabled FATS associated with AChE inhibitors, convulsants, narcotics, respiratory blockers, respiratory membrane irritants, and uncouplers to be correctly predicted.

This article considers the problem of educational placement. Several discriminant techniques are applied to a data set from a survey project of science ability. A profile vector for each student consists of five science-educational indictors. The students are intended to be placed into three reference groups: advanced, regular, and remedial. Various discriminant techniques, including Fisher’s discriminant analysis and kernel-based nonparametric discriminant analysis, are compared. The evaluation work is based on the leaving-one-out misclassification score. Results from the five school data sets and 500 bootstrap samples reveal that the kernel-based nonparametric approach with bandwidth selected by cross validation performs reasonably well. The authors regard kernel-based nonparametric procedures as desirable competitors to Fisher’s discriminant rule for handling problems of educational placement.

We present a novel action recognition method based on space-time locally adaptive regression kernels and the matrix cosine similarity measure. The proposed method uses a single example of an action to find similar matches. It does not require prior knowledge about actions; foreground/background segmentation, or any motion estimation or tracking. Our method is based on the computation of novel space-time descriptors from a query video, which measure the likeness of a voxel to its surroundings. Salient features are extracted from said descriptors and compared against analogous features from the target video. This comparison is done using a matrix generalization of the cosine similarity measure. The algorithm yields a scalar resemblance volume, with each voxel indicating the likelihood of similarity between the query video and all cubes in the target video. Using nonparametric significance tests and non-maxima suppression, we detect the presence and location of actions similar to the query video. High performance is demonstrated on challenging sets of action data containing fast motions, varied contexts, and even when multiple complex actions occur simultaneously within the field of view. Further experiments on the Weizmann and KTH datasets demonstrate state-of-the-art performance in action categorization, despite the use of only a single example.

aims to vectorize maps [1]. At first, we were restricted to topographic maps, afterwards we got interested in other kind of maps (architectural, electrical...). We include a brief explanation of the global system and then focus on the problem of locating and recognizing characters present on the map. For character location, we use some information from the previous stages of the vectorizing process (line information); we model characters as sets of short and nearby lines. We use a neural network to recognize the characters. The paper presents all the algorithms and the results we have obtained.

I am grateful for the privilege of working in the lab of my advisor, Dr. Heang-Ping Chan, along with co-advisors Dr. Berkman Sahiner and Dr. Lubomir Hadjiiski. Under their tutelage and personal attention, I have grown as a researcher, learning to read papers critically, design experiments and interpret results, and explore ways to advance the field of computer-aided diagnosis. In addition, they have provided valuable guidance on the development of computer vision techniques for CAD. It is difficult to list all the ways they have been there for me in my development in such a few short sentences. I am also thankful for my Engineering co-advisor, Dr. Jeff Fessler, as he has continued to provide advice and discussion. From that first academic advising meeting, to discussions about the research, to presentations at group meetings, he has been a source of encouragement and understanding. I enjoyed taking the medical imaging and image processing courses with Dr. Doug Noll, whose ultrasound assignment turned out to be an image of a smiley face. I also thank Dr. Chuck Meyer for his image processing expertise, enthusiasm, and support in serving on this dissertation committee and previous master’s thesis and qualifying exam committees.

We present a novel approach to change detection between two brain MRI scans (reference and target.) The proposed method uses a single modality to find subtle changes; and does not require prior knowledge (learning) of the type of changes to be sought. The method is based on the computation of a local kernel from the reference image, which measures the likeness of a pixel to its surroundings. This kernel is then used as a feature and compared against analogous features from the target image. This comparison is made using cosine similarity. The overall algorithm yields a scalar dissimilarity map (DM), indicating the local statistical likelihood of dissimilarity between the reference and target images. DM values exceeding a threshold then identify meaningful and relevant changes. The proposed method is robust to various challenging conditions including unequal signal strength. Index Terms — change detection, magnetic resonance imaging (MRI), local regression kernel 1.