This PDF file contains the editorial “Special Section on Internet Imaging” for JEI Vol. 11 Issue 04

Although many three-dimensional (3D) medical imaging visualization methods exist, 3D volume slicing remains the most commonly used technique for visualizing medical data from modalities such as CT, MRI, and PET. We propose to extend the possibilities of oblique slicing to developable curved surfaces that can be flattened and displayed in two dimensions without deformation. Such surfaces can be used to follow curved anatomical structures while preserving distance metrics at visualization time. They may also be useful for the staging of tumors, i.e., to evaluate the spatial extension of a tumor. We propose an out of core algorithm that runs in parallel on a multi-PC architecture and is able to extract surfaces from very large 3D datasets such as the visible human data set (man: 13 GB, woman: 49 GB). Experimental performance results are presented which demonstrate that parallel surface extraction is scalable and has a reasonable overhead compared with traditional oblique planar slicing. Surface extraction is made available to the public as one of the services offered by EPFLs visible human web server (http://visiblehuman.epfl.ch).

This paper describes a database collecting natural color images, called the Natural Image Database, and shows how it is used for illuminant estimation problems. First, we present significant features of the database. This database is not only a collection of color images of natural scenes, but also includes scene illuminant spectra and camera calibration data. Because these data sets are available via the Internet at http://www.osakac.ac.jp/labs/shoji, the database can be used as the calibrated common test data for evaluating the performance of illuminant estimation algorithms. Second, we focus on two illuminant estimation methods, the subspace method for estimating scene illuminant spectra and the sensor correlation method for classifying illuminant color temperature. The algorithms are briefly described. The former is based on the linear model representation of possible illuminant spectra, and the latter is based on correlation computation between an observed image and the reference illuminant gamuts. The performance of these algorithms is evaluated using the database.

Numerous research works about the extraction of low-level features from images and videos have been published. However, only recently the focus has shifted to exploiting low-level features to classify images and videos automatically into semantically broad and meaningful categories. In this paper, novel classification algorithms are presented for three broad and general-purpose categories. In detail, we present algorithms for distinguishing photo-like images from graphical images, actual photos from only photo-like, but artificial images and presentation slides/scientific posters from comics. On a large image database, our classification algorithm achieved an accuracy of 97.69% in separating photo-like images from graphical images. In the subset of photo-like images, true photos could be separated from ray-traced/rendered image with an accuracy of 97.3%, while with an accuracy of 99.5% the subset of graphical images was successfully partitioned into presentation slides/scientific posters and comics.

This paper presents a data model for images immersed in the world wide web and that derive their meaning from visual similarity, from the connection with the text of the pages that contain them, and from the link structure of the web. I will model images on the web as a graph whose nodes are either text documents or images, and whose edges are links, labeled with measures of relevance of one document towards the other. The paper presents briefly the features used to characterize the text and the visual aspect of the images, and then goes on to present a data algebra suitable to navigate and query the database.

A new pipelined architecture for colonoscopic image segmentation based on Otsu thresholding and fuzzy region growing techniques has been proposed in this paper. The computed center of mass and the segmented colon lumen of interest are used to guide the colonoscope to maneuver smoothly in the highly flexible colon with many bends, twists, and pockets. Unlike image segmentation used for colon abnormalities analysis and disease status assessment, automation of colonoscope navigation requires processing speed often unmatchable by software codes running on a general-purpose computer. In this paper, a fuzzy region growing based segmentation process has been implemented on a field programable gate array to facilitate real time computation of the lumen region. The proposed design is control path dominant and occupies less than 100k logic gates. Experimental results show that display rate of 50 frames/s. is achievable for standard colonoscopic images of size 256×256.

Efficient edge operators such as those based on partial derivatives fail to return continuous edge maps. To address this, a supplementary edge linking step is required to complete initial edge information. In this paper we propose a fast and efficient algorithm for edge linking using local information around edge terminators. In order to minimize incorrect linking decisions, the direction and the linking path for each edge terminator are established by minimizing a cost function. The particular novelty of this approach lies in the labeling scheme which assigns the directionality of the edge terminators (endpoints) based only on local knowledge. As a consequence, it relaxes the demand of a priori knowledge and furthermore assures an accurate and efficient search for edge paths in the image.

We propose a new and fast full search (FS) motion estimation algorithm to reduce the amount of computation of the FS algorithm. Computational reduction of our FS motion estimation algorithm comes from fast elimination of unlikely motion vectors. We obtain faster elimination of inappropriate motion vectors using efficient matching units by localization of a complex area in the image data. In this paper, we show three properties in block matching of motion estimation. We suggest two fast matching scan algorithms, one from an adaptive matching scan and the other from fixed dithering order. The experiments show that our scheme reduces computations by about 30% compared with conventional fast FS algorithms.

This paper presents the "Volume-Matcher 3D" project, an approach for a data-driven comparison and registration of three-dimensional (3D) images. The approach is based on a neural network model derived from self-organizing maps and extended in order to match a full 3D data set of a "source volume" with the 3D data set of a "target volume." The method developed has been tested on real cases of interest in medical imaging. The results have been evaluated on the basis of both an objective mathematical function and visual analysis performed by an expert. The software was implemented on a high performance PC using AVS/Express^TM.

This paper presents a morphology-based method for extracting license plates from cluttered images. The proposed system consists of three major components. First, a morphology-based method is proposed for extracting important contrast features as guides to search for desired license plates. The contrast feature is robust to lighting changes and invariant to different transformations like image scaling, translation, and skewing. Then, a recovery algorithm is applied to reconstruct a license plate if it is fragmented into pieces. The last stage of this method is to do license plate verification. The criterion for verification is based on the number of characters appearing in the plate that can be extracted from a clustering algorithm. The morphology-based method can significantly reduce the number of possible characters extracted and thus speeds up subsequent plate recognition. Since the feature extracted is robust to different image changes, the proposed method works well in extracting differently illuminated and oriented license plates. The average accuracy of detection is 98%. Due to the simplicity of the proposed method, all the license plates can be extracted very fast (in less than 0.5 s). The experimental results show that the proposed method improves the state-of-the-art work in terms of effectiveness and robustness for license plate detection.

In this paper, we propose color design criteria for uniform color textures, and present a new colorant-based color halftoning technique which meets these criteria, and hence generates high quality color halftones. The basic idea behind our halftoning method is to control the quality of each colorant texture separately along with the total dot distribution. In order to achieve this, we first set the total dot arrangement and then color the dots optimally without altering the total dot arrangement. This corresponds to a constrained optimization problem, and we solve this via the swap-only direct binary search heuristic. We employ a luminance human visual system model to enforce exclusion of visually nonhomogeneous patterns for the total dot distribution, as well as for each individual colorant texture. Examples of color halftone images generated using our technique are compared to images produced by a plane-independent color halftoning algorithm and by the HP 970 Cx inkjet printer driver.

This PDF file contains the errata for “JEI Vol. 11 Issue 04 Paper 1506930” for JEI Vol. 11 Issue 04

This PDF file contains the editorial “Exploratory Image Databases: Content-Based Retrieval” for JEI Vol. 11 Issue 04