https://ai.stanford.edu/~ronnyk/wrappersPrint.pdf

Wrappers for feature subset selection

A large number of algorithms have been proposed for feature subset selection. Our experimental results show that the sequential forward floating selection algorithm, proposed by Pudil et al. (1994), dominates the other algorithms tested. We study the problem of choosing an optimal feature set for land use classification based on SAR satellite images using four different texture models. Pooling features derived from different texture models, followed by a feature selection results in a substantial improvement in the classification accuracy. We also illustrate the dangers of using feature selection in small sample size situations.

/pdf/feature-selection-evaluation-application-and-small-sample-4zbdu7rpee.pdf

Feature selection: evaluation, application, and small sample performance

(1991). Applied Linear Regression Models. Journal of Quality Technology: Vol. 23, No. 1, pp. 76-77.

Applied Linear Regression Models

We present a novel image operator that seeks to find the value of stroke width for each image pixel, and demonstrate its use on the task of text detection in natural images. The suggested operator is local and data dependent, which makes it fast and robust enough to eliminate the need for multi-scale computation or scanning windows. Extensive testing shows that the suggested scheme outperforms the latest published algorithms. Its simplicity allows the algorithm to detect texts in many fonts and languages.

/pdf/detecting-text-in-natural-scenes-with-stroke-width-transform-1p4jkf3lrz.pdf

Detecting text in natural scenes with stroke width transform

Comparison of algorithms that select features for pattern classifiers

Automatic text location (without character recognition capabilities) deals with extracting image regions that contain text only. The images of these regions can then be fed to an optical character recognition module or highlighted for users. This is very useful in a number of applications such as database indexing and converting paper documents to their electronic versions. The performance of our automatic text location algorithm is shown in several applications. Compared with some traditional text location methods, our method has the following advantages: 1) low computational cost; 2) robust to font size; and 3) high accuracy.

Automatic text location in images and video frames

Transforming a paper document to its electronic version in a form suitable for efficient storage, retrieval, and interpretation continues to be a challenging problem. An efficient representation scheme for document images is necessary to solve this problem. Document representation involves techniques of thresholding, skew detection, geometric layout analysis, and logical layout analysis. The derived representation can then be used in document storage and retrieval. Page segmentation is an important stage in representing document images obtained by scanning journal pages. The performance of a document understanding system greatly depends on the correctness of page segmentation and labeling of different regions such as text, tables, images, drawings, and rulers. We use the traditional bottom-up approach based on the connected component extraction to efficiently implement page segmentation and region identification. A new document model which preserves top-down generation information is proposed based on which a document is logically represented for interactive editing, storage, retrieval, transfer, and logical analysis. Our algorithm has a high accuracy and takes approximately 1.4 seconds on a SGI Indy workstation for model creation, including orientation estimation, segmentation, and labeling (text, table, image, drawing, and ruler) for a 2550/spl times/3300 image of a typical journal page scanned at 300 dpi. This method is applicable to documents from various technical journals and can accommodate moderate amounts of skew and noise.

Document representation and its application to page decomposition

A robust and fast skew detection algorithm for generic documents

A major challenge in microbial ecology is to develop reliable and facile methods of computer-assisted microscopy that can analyze digital images of complex microbial communities at single cell resolution, and compute useful quantitative characteristics of their organization and structure without cultivation. Here we describe a computer-aided interactive system to analyze the high degree of morphological diversity in growing microbial communities revealed by phase-contrast microscopy. The system, called "CMEIAS" (Center for Microbial Ecology Image Analysis System) consists of several custom plug-ins for UTHSCSA ImageTool, a free downloadable image analysis program operating on a personal computer in a Windows NT environment. CMEIAS uses various measurement features and two object classifiers to extract size and shape measurements of segmented, digital images of microorganisms and classify them into their appropriate morphotype. The first object classifier uses a single measurement feature to analyze relatively simple communities containing only a few morphotypes (e.g., regular rods, cocci, filaments). A second new hierarchical tree classifier uses an optimized subset of multiple measurement features to analyze significantly more complex communities containing greater morphological diversity than ever before possible. This CMEIAS shape classifier automatically categorizes each cell into one of 11 predominant bacterial morphotypes, including cocci, spirals, curved rods, U-shaped rods, regular straight rods, unbranched filaments, ellipsoids, clubs, rods with extended prostheca, rudimentary branched rods, and branched filaments. The training and testing images for development and evaluation of the CMEIAS classifier were obtained from 1,937 phase-contrast grayscale digital images of various diverse communities. The CMEIAS shape classifier had an accuracy of 96.0% on a training set of 1,471 cells and 97.0% on a test set of 4,270 cells representing all 11 bacterial morphotype classes, indicating that accurate classification of rich morphological diversity in microbial communities is now possible. An interactive edit feature was added to address the main sources of error in automatic shape classification, enabling the operator to inspect the assigned morphotype of each bacterium based on visual recognition of its distinctive pseudocolor, reassign it to another morphotype class if necessary, and add up to five other morphotypes to the classification scheme. The shape classifier reports on the number and types of different morphotypes present and the abundance among each of them, thus providing the data needed to compute the morphological diversity within the microbial community. An example of how CMEIAS can augment the analysis of microbial community structure is illustrated by studies of morphological diversity as an indicator of dynamic ecological succession following a nutrient shift-up perturbation in two continuously fed, anaerobic bioreactors with morphologically distinct start communities. Various steps to minimize the limitations of computer-assisted microscopy to classify bacterial morphotypes using CMEIAS are described. In summary, CMEIAS is an accurate, robust, flexible semiautomatic computing tool that can significantly enhance the ability to quantitate bacterial morphotype diversity and should serve as a useful adjunct to the analysis of microbial community structure. This first version of CMEIAS will be released as free, downloadable plug-ins so it can provide wide application in studies of microbial ecology.

CMEIAS: A Computer-Aided System for the Image Analysis of Bacterial Morphotypes in Microbial Communities

Lane boundary detection is the problem of estimating the geometric structure of the lane boundaries of a road based on the images grabbed by a camera on board a vehicle. We use the Hough transform to detect lane boundaries with a parabolic model under a variety of road pavement types, lane structures and weather conditions. In the three-dimensional Hough space, a parabolic curve is represented as a straight line. To simplify the computation, the parametric space can be divided into (i) a two-dimensional space measured by the parameters which are shared by all the lane edges, and (ii) a one-dimensional space of the parameter which makes a distinction among different edges in an image. A multiresolution strategy is used to improve both the speed and accuracy of the Hough transform. Experimental results show that the proposed method is relatively less prone to the image noise and is computationally tractable.

/pdf/lane-boundary-detection-using-a-multiresolution-hough-5eunf7qlfb.pdf

Bin Yu

Papers

Automatic text location in images and video frames

Document representation and its application to page decomposition

A robust and fast skew detection algorithm for generic documents

CMEIAS: A Computer-Aided System for the Image Analysis of Bacterial Morphotypes in Microbial Communities

Lane boundary detection using a multiresolution Hough transform