Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Computerized Medical Imaging and Graphics 34 (2010) 415–417 Contents lists available at ScienceDirect Computerized Medical Imaging and Graphics journal homepage: www.elsevier.com/locate/compmedimag Guest editorial Introduction to the special issue on biomedical image technologies and methods 1. Introduction Today, medical technology news more and more often reports on recent developments in the field of biomedical image technologies and systems able to detect the most subtle changes in the human body and prepared to provide early diagnoses and tools for efficient treatment management and planning. It is indeed a fact that medical and molecular imaging today drives advancements in healthcare technologies, leading the way towards the concept of the personalized medicine that is gaining impulse nowadays. Non-invasive (or at least minimally invasive) imaging procedures allow physicians to evaluate a patient’s condition within minutes and take decisions on the proper therapy even in emergency cases, such like myocardial infarctions, etc. It is true that virtually all significant medical acts today involve imaging procedures, either for diagnosis, treatment planning or therapy follow-up: imaging technologies now form an inseparable element of modern medicine, which would hardly advance without their assistance. The aim of this special issue is to provide a snapshot of emerging technologies and methods in biomedical imaging, based on a selected number of papers presented in the 8th IEEE International Conference on Bio-Informatics and Bio-Engineering (BIBE 2008). The best scoring contributions to the Conference during the peer evaluation process in the tracks of Bioimaging and Biomedical Image Processing have been invited to submit extended and updated papers for publication in this special issue. A total of 21 works were initially invited and 19 papers have been submitted, although one of them has been somehow deviated and published in an earlier issue of this journal. Five papers have been considered as inappropriate for publication in this special issue after the peer review process, during which at least two reviewers (and up to four in some cases) have anonymously evaluated each submission. Of the 13 papers finally included in this special issue, 7 works have required two review rounds before their acceptance for publication. 2. Paper summaries A novel method for unsupervised DNA microarray gridding based on Support Vector Machines (SVMs) is presented in the work by Bariamis et al. [1], in their attempt to provide a robust separation method of the spots visible in DNA microarrays into distinct cells. The proposed method has been evaluated on reference microarray images containing more than two million spots in total and the researchers have proven its robustness in the presence of artifacts, noise and weakly expressed spots, as well as image rotation. 0895-6111/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.compmedimag.2010.06.001 Paul et al. [2] propose a new and simple method to quantify the signal dependent and non-signal dependent components of the noise from a single fluorescence microscopy image, without the need to use a reference image. Moreover, their method is computationally efficient and permits the on-line quantification of the noise in the acquired images. The paper by Gorpas et al. [3] proposes a novel method to “see” a tumor embedded in tissue, which is a turbid medium, in fluorescence molecular imaging. This problem has not been fully encountered yet, due to the non-linear nature of the inverse problem. Through the proposed methodology, the results of a forward solver based on a region of interest are given when comparing the simulated and the acquired data. With this modus operandi a-priori information about the initial fluorophore distribution becomes available, leading to a more feasible inverse problem solution. A system for integrated analysis of fluorescence micrographs of tissue samples is presented by Herold et al. [4]. The system combines image segmentation and information visualization principles and is proved to achieve high accuracy evaluation of the images and considerably speeds up the process. The problem of nuclear segmentation in cancer tissue imaging that is critical for specific protein activity quantification and for cancer diagnosis and therapy is investigated by Cataldo et al. [5]. A fully automated morphology-based technique able to perform accurate nuclear segmentations in images with heterogeneous staining and multiple tissue layers is presented. The proposed method gave better results when it was compared with a semi-automated method based on active contours on the segmentation of immunohistochemical images. The paper by Al-Diri et al. [6] presents an algorithm that forms a retinal vessel graph. The network connectivity is based on cost functions modeled through self-organizing feature maps (SOFMs), while overlapping vessels are handled with specialized algorithms. The algorithm is tested on junctions drawn from the public-domain DRIVE database. Iakovidis et al. [7] propose a novel methodology for screening of the small intestine with Wireless Capsule Endoscopy (WCE), which is a revolutionary, relatively non-invasive imaging technique performed by a wireless swallowable endoscopic capsule transmitting thousands of video frames per examination. In this work a data reduction algorithm is applied according to a novel scheme for the extraction of representative video frames from a full length WCE video. The technique has been experimentally validated, demonstrating that a reduction up to 85% in the reading times can be achieved for the clinical examiners. An integrated and interactive decision support system for the automated melanoma recognition of the dermoscopic images 416 Guest editorial / Computerized Medical Imaging and Graphics 34 (2010) 415–417 based on image retrieval by content and multiple expert fusion is presented in Ref. [8]. The goal of this work is to support decisionmaking by retrieving and displaying relevant past cases as well as predicting the image categories (e.g., melanoma, benign and dysplastic nevi) by combining outputs from different classifiers. A thresholding-based segmentation method is employed and the lesion-specific local color and texture features are extracted for classification. The method is shown to be effective and might constitute the basis of a clinical application operating in real time. Size-adapted segmentation of individual microcalcifications in mammography based on microcalcification scale-space signature estimation of multiscale active contours is presented in the paper by Arikidis et al. [9]. The proposed method achieved an area overlap of 0.61 ± 0.15 (when compared to manual delineations), statistically outperforming two other segmentation methods it was compared to. Moreover, the proposed method performed equally well for both small (<460 m) and large (≥460 m) microcalcifications. Noise in medical images reduces the effectiveness of texture analysis and the accuracy of medical diagnosis. In Ref. [10] seven texture measurement methods are applied to computed tomography (CT) lung tumor images under noisy conditions. The susceptibility of each texture measure to noise is investigated, aiming to enable physicians to use an automated lung texture classification system in order to choose the appropriate feature extraction algorithm. Low dose levels are used in dedicated breast CT, resulting in considerable noise levels, compared to conventional mammography. A new fuzzy distributed algorithm for assessing white matter connectivity in the brain using diffusion-weighted magnetic resonance images is presented in Ref. [11]. Fuzzy connectedness tractography can use any number of fiber orientations per voxel and can therefore be combined with any model-based or modelfree reconstruction technique that provides such estimates. In this work, Sotiropoulos et al. illustrate the performance of the method using diffusion tensor and Q-ball imaging and assess its flexibility in the way neighboring voxel similarity is assessed, thus allowing the incorporation of different affinity functions and image modalities that may even be application-specific. A new method for motion estimation of tagged cardiac magnetic resonance sequences based on variational techniques is presented by Carranza et al. [12]. The variational method has been improved by adding a new term in the optical flow equation that incorporates tracking points with high stability of phase. Results were obtained through simulated and real data, and were validated by manual tracking and with respect to a reference state-of-the-art method: harmonic phase imaging (HARP). The error, measured in pixels per frame, obtained with the proposed variational method is one order of magnitude smaller than the one achieved by the reference method, and it requires a lower computational cost. Wang et al. [13] present a novel automated monocular video monitoring approach to recover the human pose in conditions with persistently heavy obscuration, in their effort to study sleep apnoea. They employ video monitoring to analyze the covered human body during sleep in order to analyze postural changes and activities, like periodic limb movements during sleep that might have diagnostic value for sleep apnoea and other conditions. 3. Concluding remarks Biomedical imaging will undoubtedly constitute one indispensable element towards the current vision in healthcare for personalized, preventive, and cost-effective medicine. In order to materialize this vision, the joint coordinated development of research areas at multilevel scale, from molecular and cellular biological imaging up to the clinical and epidemiological level in medical imaging is considered of utmost importance. To serve this goal, a coordinated and homogeneous program for the deployment of imaging infrastructures is necessary, as well as the wide provision of access and continuous training to state-of-the-art medical and molecular imaging instrumentation and technologies, and their sustained development via basic research, multidisciplinary projects and related activities. Bold initiatives towards the above mentioned goals have already been initiated worldwide. The European Institute for Biomedical Imaging Research (EIBIR, www.eibir.org) has been recently established in order to support networking activities in research, and to spread good practice and promote common initiatives and interoperability in the field of biomedical imaging research. The aim is to generate critical mass and help coordinate research into new instrumentation, new methods, concepts and technologies that would eventually develop the field further and lead to improved diagnosis, and treatment, as well as prevention of disease. On the other hand, a joint initiative of EIBIR with the European Molecular Biology Laboratory (EMBL) has created a roadmap of the European Strategy Forum for Research Infrastructures (ESFRI) towards the establishment of the European Research Infrastructure for Imaging Technologies in Biological and Biomedical Sciences (EuroBioImaging, www.eurobioimaging.eu). The papers in this special issue have covered a wide range of theoretical developments, methodological approaches and concrete innovative applications, demonstrating the promising potential of biomedical imaging technologies and methods for the field of modern healthcare. Acknowledgements The guest editors wish to thank all the participants in the IEEE BIBE 2008, which led to this special issue, and especially, the authors of this special issue for contributing such high quality papers. We would also like to thank the referees who have critically evaluated the papers. References [1] Bariamis D, Maroulis D, Iakovidis DK. Unsupervised SVM-based gridding for DNA microarray images. Comput Med Imag Graph 2010;34:418–25. [2] Paul P, Duessmann H, Bernas T, Huber H, Kalamatianos D. Automatic noise quantification for confocal fluorescence microscopy images. Comput Med Imag Graph 2010;34:426–34. [3] Gorpas D, Yova D, Politopoulos K. A new method for processing the forward solver data in fluorescence molecular imaging. Comput Med Imag Graph 2010;34:435–45. [4] Herold J, Zhou L, Abouna SS, Pelengaris S, Epstein D, Khan M, et al. Integrating automated semantic annotation and information visualization for the analysis of multichannel fluorescence micrographs from pancreatic tissue. Comput Med Imag Graph 2010;34:446–52. [5] Cataldo SD, Ficarra E, Acquaviva A, Macii E. Achieving the way for automated segmentation of nuclei in cancer tissue images through morphologybased approach: a quantitative evaluation. Comput Med Imag Graph 2010;34:453–61. [6] Al-Diri BI, Hunter A, Steel D, Habib M. Automated analysis of retinal vascular network connectivity. Comput Med Imag Graph 2010;34:462–70. [7] Iakovidis D, Tsevas S, Polydorou A. Reduction of capsule endoscopy reading times by unsupervised image mining. Comput Med Imag Graph 2010;34:471–8. [8] Rahman M, Bhattacharya P. An integrated and interactive decision support system for automated melanoma recognition of dermoscopic images. Comput Med Imag Graph 2010;34:479–86. [9] Arikidis NS, Karahaliou A, Skiadopoulos S, Korfiatis P, Likaki E, Panayiotakis G, et al. Size-adapted microcalcification segmentation in mammography utilizing scale-space signatures. Comput Med Imag Graph 2010;34:487–93. [10] Al-Kadi OS. Assessment of texture measures susceptibility to noise in conventional and contrast enhanced computed tomography lung tumour images. Comput Med Imag Graph 2010;34:494–503. [11] Sotiropoulos SN, Bai L, Tench CR. Fuzzy anatomical connectedness of the brain using single and multiple fibre orientations estimated from diffusion MRI. Comput Med Imag Graph 2010;34:504–13. Guest editorial / Computerized Medical Imaging and Graphics 34 (2010) 415–417 [12] Carranza N, Bajo A, Sroubek F, Santamarta C, Cristobal G, Santos A, et al. Motion estimation of tagged cardiac magnetic resonance images using variational techniques. Comput Med Imag Graph 2010;34:514–22. [13] Wang C-W, Hunter A, Gravill N, Matusiewicz S. Real time pose recognition of covered human for diagnosis of sleep apnoea. Comput Med Imag Graph 2010;34:523–33. Guest Managing Editor George Kontaxakis ∗ E.T.S.I. Telecomunicación, Universidad Politécnica de Madrid, Spain Guest Editor Nikolaos Bourbakis Engineering College, Wright State University, OH, USA 417 Guest Editor Konstantina S. Nikita Department of Electrical and Computer Engineering, National Technical University of Athens, Greece Guest Editor Constantinos S. Pattichis Department of Computer Science, University of Cyprus, Cyprus ∗ Corresponding author. Tel.: +34 91 336 7366/4253; fax: +34 91 337 7323. E-mail address: [email protected] (G. Kontaxakis)