Artif Intell Med.

Automatic image-based assessment of lesion development during hemangioma follow-up examinations.

Publication Date: 2010 Aug 20 PMID: 20729044
Authors: Zambanini, S. - Sablatnig, R. - Maier, H. - Langs, G.
Journal: Artif Intell Med

OBJECTIVE: This paper presents an automatic method for the quantification of the development of cutaneous hemangiomas in digital images. Two measurements on digital images acquired during follow-up examinations are performed: (1) the skin area affected by the lesion is measured and (2) the change of the hemangioma during follow-up examinations called regression is determined. Current manual measurements exhibit inter- and intra-reader variation, which impedes precision and comparisons across clinical studies. The proposed automatic method aims at a more accurate and objective evaluation of the course of disease than the current clinical practice of manual measurement. METHODS AND MATERIAL: The proposed method classifies individual pixels and calculates the area based on a ruler attached to the skin. For the regression detection follow-up images are registered automatically based on local gradient histograms. The method was evaluated on 90 individual images and a set of 4 follow-up series consisting of 3-4 examinations. RESULTS: The absolute average error of the individual area measurements lies at 0.0775cm(2) corresponding to a variation coefficient of 8.82%. The measurement of the regression area provides an absolute average error of 0.1134cm(2) and a variation coefficient of 7.40 %. CONCLUSIONS: The results indicate that the proposed method provides an accurate and objective evaluation of the course of cutaneous hemangiomas. This is relevant for the monitoring of individual therapy and for clinical trials.

post to: CiteULike

Knowledge discovery and computer-based decision support in biomedicine.

Publication Date: 2010 Sep PMID: 20615677
Authors: Soda, P. - Pechenizkiy, M. - Tortorella, F. - Tsymbal, A.
Journal: Artif Intell Med

post to: CiteULike

Computer-aided diagnosis of pulmonary nodules using a two-step approach for feature selection and classifier ensemble construction.

Publication Date: 2010 Sep PMID: 20570118
Authors: Lee, M. C. - Boroczky, L. - Sungur-Stasik, K. - Cann, A. D. - Borczuk, A. C. - Kawut, S. M. - Powell, C. A.
Journal: Artif Intell Med

OBJECTIVE: Accurate classification methods are critical in computer-aided diagnosis (CADx) and other clinical decision support systems. Previous research has reported on methods for combining genetic algorithm (GA) feature selection with ensemble classifier systems in an effort to increase classification accuracy. In this study, we describe a CADx system for pulmonary nodules using a two-step supervised learning system combining a GA with the random subspace method (RSM), with the aim of exploring algorithm design parameters and demonstrating improved classification performance over either the GA or RSM-based ensembles alone. METHODS AND MATERIALS: We used a retrospective database of 125 pulmonary nodules (63 benign; 62 malignant) with CT volumes and clinical history. A total of 216 features were derived from the segmented image data and clinical history. Ensemble classifiers using RSM or GA-based feature selection were constructed and tested via leave-one-out validation with feature selection and classifier training executed within each iteration. We further tested a two-step approach using a GA ensemble to first assess the relevance of the features, and then using this information to control feature selection during a subsequent RSM step. The base classification was performed using linear discriminant analysis (LDA). RESULTS: The RSM classifier alone achieved a maximum leave-one-out Az of 0.866 (95% confidence interval: 0.794-0.919) at a subset size of s=36 features. The GA ensemble yielded an Az of 0.851 (0.775-0.907). The proposed two-step algorithm produced a maximum Az value of 0.889 (0.823-0.936) when the GA ensemble was used to completely remove less relevant features from the second RSM step, with similar results obtained when the GA-LDA results were used to reduce but not eliminate the occurrence of certain features. After accounting for correlations in the data, the leave-one-out Az in the two-step method was significantly higher than in the RSM and the GA-LDA. CONCLUSIONS: We have developed a CADx system for evaluation of pulmonary nodule based on a two-step feature selection and ensemble classifier algorithm. We have shown that by combining classifier ensemble algorithms in this two-step manner, it is possible to predict the malignancy for solitary pulmonary nodules with a performance exceeding that of either of the individual steps.

post to: CiteULike

Fusing visual and clinical information for lung tissue classification in high-resolution computed tomography.

Publication Date: 2010 Sep PMID: 20547044
Authors: Depeursinge, A. - Racoceanu, D. - Iavindrasana, J. - Cohen, G. - Platon, A. - Poletti, P. A. - Muller, H.
Journal: Artif Intell Med

OBJECTIVE: We investigate the influence of the clinical context of high-resolution computed tomography (HRCT) images of the chest on tissue classification. METHODS AND MATERIALS: 2D regions of interest in HRCT axial slices from patients affected with an interstitial lung disease are automatically classified into five classes of lung tissue. Relevance of the clinical parameters is studied before fusing them with visual attributes. Two multimedia fusion techniques are compared: early versus late fusion. Early fusion concatenates features in one single vector, yielding a true multimedia feature space. Late fusion consisting of the combination of the probability outputs of two support vector machines. RESULTS AND CONCLUSION: The late fusion scheme allowed a maximum of 84% correct predictions of testing instances among the five classes of lung tissue. This represents a significant improvement of 10% compared to a pure visual-based classification. Moreover, the late fusion scheme showed high robustness to the number of clinical parameters used, which suggests that it is appropriate for mining clinical attributes with missing values in clinical routine.

post to: CiteULike

A segmentation framework for abdominal organs from CT scans.

Publication Date: 2010 Sep PMID: 20542673
Authors: Campadelli, P. - Casiraghi, E. - Pratissoli, S.
Journal: Artif Intell Med

OBJECTIVE: Computed tomography images are becoming an invaluable mean for abdominal organ investigation. In the field of medical image processing, some of the current interests are the automatic diagnosis of liver, spleen, and kidney pathologies, and the 3D volume rendering of these abdominal organs. Their automatic segmentation is the first and fundamental step in all these studies, but it is still an open problem. METHODS: In this paper we propose a fully automatic, gray-level based segmentation framework based on a multiplanar fast marching method. The proposed segmentation scheme is general, and employs only established and not critical anatomical knowledge. For this reason, it can be easily adapted to segment different abdominal organs, by overcoming problems due to the high inter- and intra-patient gray-level, and shape variabilities; the extracted volumes are then combined to produce the final results. RESULTS: The system has been evaluated by computing the symmetric volume overlap (SVO) between the automatically segmented (liver and spleen) volumes and the volumes manually traced by radiological experts. The test dataset is composed of 60 images, where 40 images belong to a private dataset, and 20 images to a public one. Liver segmentation has achieved an average SVO congruent with94, which is comparable to the mean intra- and inter-personal variation (96). Spleen segmentation achieves similar, promising results (SVO congruent with93). The comparison of these results with those achieved by active contour models (SVO congruent with90), and topology adaptive snakes (SVO congruent with92) proves the efficacy of our system. CONCLUSIONS: The described segmentation method is a general framework that can be adapted to segment different abdominal organs, achieving promising segmentation results. It has to be noted that its performance could be further improved by incorporating shape based rules.

post to: CiteULike

Human movement onset detection from isometric force and torque measurements: a supervised pattern recognition approach.

Publication Date: 2010 Sep PMID: 20510593
Authors: Soda, P. - Mazzoleni, S. - Cavallo, G. - Guglielmelli, E. - Iannello, G.
Journal: Artif Intell Med

OBJECTIVE: Recent research has successfully introduced the application of robotics and mechatronics to functional assessment and motor therapy. Measurements of movement initiation in isometric conditions are widely used in clinical rehabilitation and their importance in functional assessment has been demonstrated for specific parts of the human body. The determination of the voluntary movement initiation time, also referred to as onset time, represents a challenging issue since the time window characterizing the movement onset is of particular relevance for the understanding of recovery mechanisms after a neurological damage. Establishing it manually as well as a troublesome task may also introduce oversight errors and loss of information. METHODS: The most commonly used methods for automatic onset time detection compare the raw signal, or some extracted measures such as its derivatives (i.e., velocity and acceleration) with a chosen threshold. However, they suffer from high variability and systematic errors because of the weakness of the signal, the abnormality of response profiles as well as the variability of movement initiation times among patients. In this paper, we introduce a technique to optimise onset detection according to each input signal. It is based on a classification system that enables us to establish which deterministic method provides the most accurate onset time on the basis of information directly derived from the raw signal. RESULTS: The approach was tested on annotated force and torque datasets. Each dataset is constituted by 768 signals acquired from eight anatomical districts in 96 patients who carried out six tasks related to common daily activities. The results show that the proposed technique improves not only on the performance achieved by each of the deterministic methods, but also on that attained by a group of clinical experts. CONCLUSIONS: The paper describes a classification system detecting the voluntary movement initiation time and adaptable to different signals. By using a set of features directly derived from raw data, we obtained promising results. Furthermore, although the technique has been developed within the scope of isometric force and torque signal analysis, it can be applied to other detection problems where several simple detectors are available.

post to: CiteULike

A computer-aided detection system for clustered microcalcifications.

Publication Date: 2010 Sep PMID: 20472412
Authors: Marrocco, C. - Molinara, M. - D'Elia, C. - Tortorella, F.
Journal: Artif Intell Med

OBJECTIVE: The aim of this paper is to describe a novel system for computer-aided detection of clusters of microcalcifications on digital mammograms. METHODS AND MATERIAL: Mammograms are first segmented by means of a tree-structured Markov random field algorithm that extracts the elementary homogeneous regions of interest. An analysis of such regions is then performed by means of a two-stage, coarse-to-fine classification based on both heuristic rules and classifier combination. In this phase, we avoid taking a decision on the single microcalcifications and forward it to the successive phase of clustering realized through a sequential approach. RESULTS: The system has been tested on a publicly available database of mammograms and compared with previous approaches. The obtained results show that the system is very effective, especially in terms of sensitivity. CONCLUSIONS: The proposed approach exhibits some remarkable advantages both in segmentation and classification phases. The segmentation phase employs an image model that reduces the computational burden, preserving the small details in the image through an adaptive local estimation of all model parameters. The classification stage combines the results of the classifiers focused on the single microcalcification and the cluster as a whole. Such an approach makes a detection system particularly effective and robust with respect to the large variations exhibited by the clusters of microcalcifications.

post to: CiteULike

Fusion of fuzzy statistical distributions for classification of thyroid ultrasound patterns.

Publication Date: 2010 Sep PMID: 20427164
Authors: Iakovidis, D. K. - Keramidas, E. G. - Maroulis, D.
Journal: Artif Intell Med

OBJECTIVE: This paper proposes a novel approach for thyroid ultrasound pattern representation. Considering that texture and echogenicity are correlated with thyroid malignancy, the proposed approach encodes these sonographic features via a noise-resistant representation. This representation is suitable for the discrimination of nodules of high malignancy risk from normal thyroid parenchyma. MATERIALS AND METHODS: The material used in this study includes a total of 250 thyroid ultrasound patterns obtained from 75 patients in Greece. The patterns are represented by fused vectors of fuzzy features. Ultrasound texture is represented by fuzzy local binary patterns, whereas echogenicity is represented by fuzzy intensity histograms. The encoded thyroid ultrasound patterns are discriminated by support vector classifiers. RESULTS: The proposed approach was comprehensively evaluated using receiver operating characteristics (ROCs). The results show that the proposed fusion scheme outperforms previous thyroid ultrasound pattern representation methods proposed in the literature. The best classification accuracy was obtained with a polynomial kernel support vector machine, and reached 97.5% as estimated by the area under the ROC curve. CONCLUSIONS: The fusion of fuzzy local binary patterns and fuzzy grey-level histogram features is more effective than the state of the art approaches for the representation of thyroid ultrasound patterns and can be effectively utilized for the detection of nodules of high malignancy risk in the context of an intelligent medical system.

post to: CiteULike