Catálogo de publicaciones - libros

In this paper, a new lossy image compression technique based on adaptive variable degree variable segment length Chebyshev polynomials is proposed. The main advantage of this method over JPEG is that it has a direct individual error control where the maximum error in gray level difference between the original and the reconstructed images can be specified by the user. This is a requirement for medical applications where near lossless quality is needed. The compression is achieved by representing the gray level variations across any determined section of a row or column of an image by the coefficients of a Chebyshev polynomial. The performance of the method was evaluated on a number of test images and using some quantitative measures compared to the well known JPEG compression techniques.

The paper presents a novel Linear Hashtable Method Predicted Hexagonal Search (LHMPHS) method for block base motion compensation. It bases on the edge motion estimation algorithm called hexagonal search (HEXBS). Most current variances of hexagonal search are investigated. On the basis of research of previous algorithms, we proposed a Linear Hashtable Motion Estimation Algorithm (LHMEA). The proposed algorithm introduces hashtable into motion estimation. It uses information from the current frame. The criterion uses spatially correlated macroblock (MB)’s information. Except for coarse search, the spatially correlated information is also used in inner search. The performance of the algorithm is evaluated by using standard video sequences and the results are compared to current algorithms such as Full Search, Logarithmic Search etc. The evaluation considers the three important metrics: time, compression rate and PSNR.

In this paper we present a method for optical paper surface roughness measurement, which overcomes the disadvantages of the traditional methods. Airflow-based roughness measurement methods and profilometer require expensive special equipment, essential laboratory conditions, are contact-based and slow and unsuitable for on-line control purposes methods. We employed an optical microscope with a built-in CCD-camera to take images of paper surface. The obtained image is considered as a texture. We applied statistical brightness measures and fractal dimension analysis for texture analysis. We have found a strong correlation between the roughness and a fractal dimension. Our method is non-contact–based, fast and is suitable for on-line control measurements in the paper industry.

We analyze multispectral reflectance images of concrete aggregate material, and design computational measures of the important and critical parameters used in concrete production. The features extracted from the images are exploited as explanatory variables in regression models and used to predict aggregate type, water content, and size distribution. We analyze and validate the methods on five representative aggregate types, commonly used in concrete production. Using cross validation, the generated models proves to have a high performance in predicting all of the critical parameters.

The Automated Multiple View Inspection (AMVI) has been recently developed for automated defect detection of manufactured objects. The approach detects defects by analysing image sequences in two steps. In the first step, potential defects are automatically identified in each image of the sequence. In the second step, the potential defects are tracked in the sequence. The key idea of this strategy is that only the existing defects (and not the false detections) can be successfully tracked in the image sequence because they are located in positions dictated by the motion of the test object. The AMVI strategy was successfully implemented for calibrated image sequences. However, it is not simple to implement it in industrial environments because the calibration process is a difficult task and unstable. In order to avoid the mentioned disadvantages, in this paper we propose a new AMVI strategy based on the tracking of potential detects in uncalibrated image sequences. Our approach tracks the potential defects based on a motion model estimated from the image sequence self. Thus, we obtain a motion model by matching structure points of the images. We show in our experimental results on aluminium die castings that the detection is promising in uncalibrated images by detecting 92.3% of all existing defects with only 0.33 false alarms per image.

Recently, a number of methods for 3-D modeling from images have been developed. However, the accuracy of a reconstructed model depends on camera positions and postures with which the images are obtained. In most of conventional methods, some skills for adequately controlling the camera movement are needed for users to obtain a good 3-D model. In this study, we propose an interactive 3-D modeling interface in which special skills are not required. This interface consists of “indication of camera movement” and “preview of reconstruction result.” In experiments for subjective evaluation, we verify the usefulness of the proposed 3D modeling interfaces.

Traditional segmentation methods cannot provide satisfying results for extraction of prostate gland from Transrectal Ultrasound (TRUS) images because of the presence of strong speckle noise and shadow artifacts. Most ultrasound image segmentation techniques that adopt model-based approach such as active contour are considered semi-automatic because they require initial seeds or contours to be manually identified. In this paper, we propose a method for automatic segmentation of prostate using feature-based self organizing map (SOM). Median filtering and top hat transform are first applied to remove speckle noise. A technique is developed to remove ultrasound-specific speckles using texture-based thresholding. An SOM algorithm is employed to identify prostate pixels taking spatial information, gray-level as well as texture information to form its input vector. The clustered image is then processed to produce a fully connected prostate contour. A number of experiments comparing extracted contours with manually-delineated contours validated the performance of our method.