Theodore L. Economopoulos

Contrast enhancement of images using Partitioned Iterated Function Systems

A new algorithm for the contrast enhancement of images, based on the theory of Partitioned Iterated Function System (PIFS), is presented. A PIFS consists of contractive transformations, such that the original image is the fixed point of the union of these transformations. Each transformation involves the contractive affine spatial transform of a square block, as well as the linear transform of the gray levels of its pixels. The transformation of the gray levels is determined by two parameters which adjust the brightness and the contrast of the transformed block. The PIFS is used in order to create a lowpass version of the original image. The contrast-enhanced image is obtained by adding the difference of the original image with its lowpass version, to the original image itself. The proposed algorithm uses a predefined constant value for the contrast parameter, whereas, the parameters of the affine spatial transform, as well as the parameter adjusting the brightness, are calculated using k-dimensional trees. The lowpass version of the original image is obtained applying the PIFS on the original image repeatedly while using a value for the contrast parameter that is lower than the predefined one. Quantitative and qualitative results stress the superior performance of the proposed contrast enhancement algorithm against four other widely used contrast enhancement methods; namely, linear and nonlinear unsharp masking, Contrast Limited Adaptive Histogram Equalization and Local Range Modification.

Volumetric difference evaluation of registered three-dimensional pre-operative and post-operative CT dental data

OBJECTIVE The purpose of this study is to propose a complete methodology for automatically registering three-dimensional (3D) pre-operative and post-operative CT scan dental volumes as well as to provide a toolset for quantifying and evaluating their volumetric differences. METHODS The proposed methodology was applied to cone beam CT (CBCT) data from 20 patients in order to assess the volume of augmented bone in the alveolar region. In each case, the pre-operative and post-operative data were registered using a 3D affine-based scheme. The performance of the 3D registration algorithm was evaluated by measuring the average distance between the edges of the registered sets. The differences between the registered sets were assessed through 3D subtraction radiography. The volume of the differences was finally evaluated by defining regions of interest in each slice of the subtracted 3D data and by combining all respective slices to model the desired volume of interest. The effectiveness of the algorithm was verified by applying it to several reference standard-shaped objects with known volumes. RESULTS Satisfactory alignment was achieved as a low average offset of 1.483 ± 1.558 mm was recorded between the edges of the registered sets. Moreover, the estimated volumes closely matched the volumes of the reference objects used for verification, as the recorded volume differences were less than 0.4 mm(3) in all cases. CONCLUSION The proposed method allows for automatic registration of 3D CBCT data sets and the volumetric assessment of their differences in particular areas of interest. The proposed approach provides accurate volumetric measurements in three dimensions, requiring minimal user interaction.

Automatic correspondence using the enhanced hexagonal centre-based inner search algorithm for point-based dental image registration

OBJECTIVES In this paper, the enhanced hexagonal centre-based inner search (EHCBIS) algorithm, for automatic point correspondence, is proposed for dental image registration. METHODS The presented algorithm is incorporated within a general registration scheme, which is based on extracting a set of candidate points on the reference image, finding their corresponding points in the image to be transformed (float image) using the proposed algorithm and applying a suitable geometrical transformation towards automatic registration. The performance of the proposed algorithm is evaluated against three well-known methods for automatic correspondence, the self-organizing maps, the automatic extraction of corresponding points and the trimmed iterative closest point method, in terms of registration accuracy. RESULTS Qualitative and quantitative results on registering 123 dental pairs show that the proposed algorithm outperforms the other methods for automatic correspondence with or without the presence of noise. CONCLUSIONS The EHCBIS method is capable of defining automatically corresponding points in dental image pairs. It can be incorporated within a general scheme for point-based registration of dental radiographs acquired with or without rigorous a priori standardization. The applied projective transformation provides a reliable model for registering intraoral radiographs. The methodology does not require any segmentation prior to alignment providing subtraction radiographs and fused images for clinical evaluation regarding the evolution of a disease or the response to a therapeutic scheme.

Geometry-based vs. intensity-based medical image registration

Spatial alignment of Computed Tomography (CT) data sets is often required in numerous medical applications and it is usually achieved by applying conventional exhaustive registration techniques, which are mainly based on the intensity of the subject data sets. Those techniques consider the full range of data points composing the data, thus negatively affecting the required processing time. Alternatively, alignment can be performed using the correspondence of extracted data points from both sets. Moreover, various geometrical characteristics of those data points can be used, instead of their chromatic properties, for uniquely characterizing each point, by forming a specific geometrical descriptor. This paper presents a comparative study reviewing variations of geometry-based, descriptor-oriented registration techniques, as well as conventional, exhaustive, intensity-based methods for aligning three-dimensional (3D) CT data pairs. In this context, three general image registration frameworks were examined: a geometry-based methodology featuring three distinct geometrical descriptors, an intensity-based methodology using three different similarity metrics, as well as the commonly used Iterative Closest Point algorithm. All techniques were applied on a total of thirty 3D CT data pairs with both known and unknown initial spatial differences. After an extensive qualitative and quantitative assessment, it was concluded that the proposed geometry-based registration framework performed similarly to the examined exhaustive registration techniques. In addition, geometry-based methods dramatically improved processing time over conventional exhaustive registration.

Automatic Correspondence on Medical Images: A Comparative Study of Four Methods for Allocating Corresponding Points

The accurate estimation of point correspondences is often required in a wide variety of medical image-processing applications. Numerous point correspondence methods have been proposed in this field, each exhibiting its own characteristics, strengths, and weaknesses. This paper presents a comprehensive comparison of four automatic methods for allocating corresponding points, namely the template-matching technique, the iterative closest points approach, the correspondence by sensitivity to movement scheme, and the self-organizing maps algorithm. Initially, the four correspondence methods are described focusing on their distinct characteristics and their parameter selection for common comparisons. The performance of the four methods is then qualitatively and quantitatively compared over a total of 132 two-dimensional image pairs divided into eight sets. The sets comprise of pairs of images obtained using controlled geometry protocols (affine and sinusoidal transforms) and pairs of images subject to unknown transformations. The four methods are statistically evaluated pairwise on all image pairs and individually in terms of specific features of merit based on the correspondence accuracy as well as the registration accuracy. After assessing these evaluation criteria for each method, it was deduced that the self-organizing maps approach outperformed in most cases the other three methods in comparison.

A contrast correction method for dental images based on histogram registration.

Contrast correction is often required in digital subtraction radiography when comparing medical data acquired over different time periods owing to dissimilarities in the acquisition process. This paper focuses on dental radiographs and introduces a novel approach for correcting the contrast in dental image pairs. The proposed method modifies the subject images by applying typical registration techniques on their histograms. The proposed histogram registration method reshapes the histograms of the two subject images in such a way that these images are matched in terms of their contrast deviation. The method was extensively tested over 4 sets of dental images, consisting of 72 registered dental image pairs with unknown contrast differences as well as 20 dental pairs with known contrast differences. The proposed method was directly compared against the well-known histogram-based contrast correction method. The two methods were qualitatively and quantitatively evaluated for all 92 available dental image pairs. The two methods were compared in terms of the contrast root mean square difference between the reference image and the corrected image in each case. The obtained results were also verified statistically using appropriate t-tests in each set. The proposed method exhibited superior performance compared with the well-established method, in terms of the contrast root mean square difference between the reference and the corrected images. After suitable statistical analysis, it was deduced that the performance advantage of the proposed approach was statistically significant.

Image contrast enhancement through regional application of partitioned iterated function systems

Abstract. A new technique is presented for enhancing the contrast in digital images, combining the theory of partitioned iterated function system (PIFS) and image segmentation. The image is first segmented through the region growing segmentation technique, and the PIFS enhancement algorithm is applied separately to each image segment. The defined PIFS of each section is modeled by a contractive transformation, which consists of an affine spatial transform, as well as the linear transform of the graylevels of image segment pixels. The transformation of the graylevels is determined by two parameters that adjust the brightness and contrast of the transformed image segment. After the PIFS algorithm is applied to each extracted image segment, a lowpass version of the original image is created. The contrast-enhanced image is obtained by suitably combining the original image with its lowpass version. The proposed regional PIFS approach was applied to numerous test images, ranging from medical data of various modalities to standard images. The obtained quantitative and qualitative results showed superior performance on behalf of the proposed method when compared with three other widely used contrast enhancement methods, namely, contrast stretching, unsharp masking, and contrast-limited adaptive histogram equalization.

Contrast enhancement of images using partitioned iterated function systems

A new algorithm for the contrast enhancement of images, based on the theory of Partitioned Iterated Function System (PIFS), is presented. A PIFS consists of contractive transformations, such that the original image is the fixed point of the union of these transformations. Each transformation involves the contractive affine spatial transform of a square block, as well as the linear transform of the gray levels of its pixels. The PIFS is used in order to create a low-pass version of the original image. The contrast-enhanced image is obtained by adding the difference of the original image with its lowpass version, to the original image itself. Quantitative and qualitative results stress the superior performance of the proposed contrast enhancement algorithm against two other widely used contrast enhancement methods.

3D dental image registration using exhaustive deformable models: a comparative study

OBJECTIVES Image registration is commonly used in dental applications for aligning imaging data sets, which is particularly useful when assessing the progression or regression of particular pathomorphic conditions. However, due to the nature of the processed data or the data acquisition process itself, rigid body registration may be insufficient to accurately align the processed data sets. In such cases, deformable models are employed. This study presents a comparison of four well-established deformable models for aligning CBCT volumes. METHODS The compared models include the original Demons algorithm, symmetric forces Demons, diffeomorphic Demons and level-set motion. The compared techniques are incorporated into a general image registration scheme featuring two distinct stages: a common, fast, rigid-based alignment for pre-registering the data and a finer elastic registration phase, based on the four compared deformation models. RESULTS The proposed framework was applied to a total of 40 CBCT volume pairs with known and unknown initial differences. CONCLUSIONS After both qualitative and quantitative assessment of the produced aligned data, it was concluded that the level-set motion method outperformed all other techniques for data pairs with both unknown initial differences, as well as with known elastic deviations based on fixed sinusoidal models and B-splines.

Alignment of three-dimensional point clouds using combined descriptors

This paper presents a new methodology for aligning three-dimensional (3D) models of objects, based on point correspondences. In this case, objects are modelled as 3D point clouds. The proposed methodology considers pairs of such point clouds and firstly down-samples them in order to further improve processing time. Then, corresponding points are allocated between the processed point clouds, by using a novel combinational descriptor scheme. Finally, a global transformation is estimated from the inliers of the obtained correspondences. This transformation is used to align the two point clouds. The proposed methodology was applied to five pairs of large scale 3D point clouds. Results indicate that the proposed scheme achieved satisfactory alignment accuracy for all tested data pairs.