SEARCH · Search PubMed

Results for “Segmentation”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 613 records · Page 34Linked to original sources

PubMed1997

Interactions between colour and motion in image segmentation.

BACKGROUND: . An important early stage in visual processing is image segmentation, in which similar regions are grouped together and segregated from dissimilar regions, so that distinct objects ultimately may be located and recognized. In the natural world, objects are simultaneously characterized by colour, motion, texture and other visual attributes. How does the human visual system combine these attributes to segment the image? Although colour and motion information are conveyed by distinct functional streams from retina to visual cortex, there is increasing evidence for early and substantial cross-talk between the streams. Here, we explore psychophysical evidence for interactions between colour and motion in image segmentation. RESULTS: . Observers performed forced-choice segmentation tasks on random-dot stimuli. The dots in the vertical target figure were distinguished from the background dots by a different distribution of speeds or colours. To explore interactions between motion and colour segmentation, we added motion noise to the colour signal (or vice versa) by assigning all dots speeds (or colours) drawn from one of several noise distributions. Motion noise severely affects segmentation by colour. Motion noise defined by a broad distribution of speeds degrades colour segmentation, but a two-speed motion distribution (half moving up, half moving down) facilitates colour segmentation. Control experiments prove that the facilitatory effect is not caused by integrating colour information over different frames, nor can it be explained by probability summation over the two planes of moving dots. Colour noise also affects motion segmentation, but under a more restricted range of conditions, and not in a facilitatory way. CONCLUSIONS: . Colour and motion information interact at early stages during image segmentation, before decisions based on either cue in isolation are made. The robust bipolar effects of motion information on segmentation by colour indicate that the establishment of motion-defined surfaces takes primacy, and that such surfaces constitute important primitives for further processing.

Choice Behavior↗

PubMed2004

Passive external cardiac constraint improves segmental left ventricular wall motion and reduces akinetic area in patients with non-ischemic dilated cardiomyopathy.

OBJECTIVE: To verify changes in left ventricular (LV) volumes and regional myocardial wall motion after implantation of a textile cardiac support device (CSD) for passive external constraint in non-ischemic dilated cardiomyopathy. METHODS: In nine male patients participating in a non-randomized clinical trial LV volumes were determined and the segmental LV wall motion was studied by contrast-enhanced electron-beam CT in a sectionwise manner at three ventricular levels (base, middle and apex of ventricle) before and 32+/-6 months after CSD implantation. In 16 myocardial segments ejection fraction and wall thickening were measured semiautomatically after drawing the myocardial contours. The wall motion score index was calculated based on semiquantitative visual grading in each segment. RESULTS: The global LV volumes decreased significantly from 304.3 +/- 90.9 to 231.5 +/- 103.9 ml at end-diastole and from 239.7 +/- 83.7 to 164.0 +/- 97.7 at end-systole (P<0.05). Overall ejection fraction increased from 14.8 +/- 8.2 to 25.7 +/- 17.1% (P<0.05). A segment-by-segment analysis demonstrated a significant increase of regional ejection fraction in the basal myocardium as well as in the mid-inferior, mid-inferolateral, and mid-anterolateral myocardium. Overall wall thickening increased from 16.4 +/- 13.3 to 24.2 +/- 18.1% (P<0.05), but without significant differences in a segment-by-segment comparison. The mean wall motion score index improved from 2.70 +/- 0.26 to 2.20 +/- 0.71 (P<0.05), with an increased wall motion in eight (89%) patients. A section-by-section analysis demonstrated significantly improved wall motion in the inferior and lateral segments at each ventricular level. Postoperatively, the number of akinetic and markedly hypokinetic segments decreased significantly (P<0.05) from 56 (39%) to 26 (18%) and from 76 (53%) to 56 (37%), respectively. CONCLUSION: CSD implantation improves segmental wall motion, predominantly in the inferior and lateral myocardium, and reduces the number of akinetic and hypokinetic segments.

Adult↗

PubMed1999

Packaging and replication regulation revealed by chimeric genome segments of double-stranded RNA bacteriophage phi6.

Bacteriophage phi6 has a double-stranded RNA genome composed of three linear segments, L, M, and S. The innermost particle in the virion of phi6, like in the other dsRNA viruses, is an RNA-dependent RNA polymerase complex, which carries out all the functions needed for the replication of the viral genome. Empty polymerase complexes can package the single-stranded copies of the viral genome segments, replicate the packaged segments into double-stranded form (minus strand synthesis), and then produce new plus strands (transcripts) from the double-stranded RNA templates. The three viral genomic segments contain unique packaging signals at their 5' ends, and minus strand synthesis initiation is dependent on the sequence at the 3' end. Here we have constructed chimeric segments that have the packaging signal from one segment and the minus strand synthesis initiation signal from another segment. Using purified recombinant polymerase complexes and single-stranded/chimeric and original RNA segments, we have analyzed the packaging and replication regulation operating in in vitro conditions. We show that the 5' end of the L genome segment in single-stranded form is needed to switch from the packaging to the minus strand synthesis and the same sequence is required in double-stranded form to switch on plus strand synthesis. In addition we have constructed deletions to the M segment to analyze the possible regulatory role of the internal noncoding area of this segment.

Bacteriophage phi 6↗

PubMed1990

Putative protein kinase product of the Drosophila segment-polarity gene zeste-white3.

The metameric pattern of the Drosophila embryo is regulated by a combination of maternal and zygotic genes. The segment-polarity class of genes are required for the correct patterning within each segmental unit. Mutations in any one of these genes results in deletions and duplications of parts of each segment. The segment-polarity genes act coordinately by means of local cellular interactions to assign and maintain an identity for each cell in the segment, and to establish segment boundaries. Here we describe the molecular characterization of a novel segment-polarity gene, zeste-white3 (zw3). Embryos derived from germ lines that are homozygous for zw3 mutations (zw3 embryos) have phenotypes similar to embryos that are mutant for the segment-polarity gene naked (nkd). These embryos lack most of the ventral denticles, which are differentiated structures derived from the most anterior region of each segment. We have isolated the zw3 gene and compared the structure of one maternal and one zygotic transcript encoded by the gene. The zw3 gene is unique among the segment-polarity genes so far characterized, in that it encodes proteins that have homology to serine-threonine protein kinases. This indicates that zw3 may play a part in a signal transduction pathway involved in the establishment of cell identity within each embryonic segment.

Amino Acid Sequence↗

PubMed1999

Effects of lovastatin on progression of non-dilated and dilated coronary segments and on restenosis in patients after PTCA. The cholesterol lowering atherosclerosis PTCA trial (CLAPT).

OBJECTIVES: The Cholesterol Lowering Atherosclerosis PTCA Trial (CLAPT) is a prospective, randomized trial with blinded angiographic end-points to assess the effect of 2-year's treatment with lovastatin initiated 4 weeks prior to PTCA, compared to usual care on non-dilated coronary segments and on dilated coronary lesions in male patients with total cholesterol between 200 and 300 mg. dl(-1)who underwent elective PTCA. METHODS AND RESULTS: Two hundred and twenty six patients were randomized 4 weeks prior to PTCA to special care (diet plus lovastatin n=112) or usual care (diet; n=114). One hundred and ninety-nine patients underwent PTCA at baseline and were finally included in the study. Quantitative coronary angiographic assessment was performed on blinded cinefilms at baseline (PTCA) and repeated after 4 and 24 months in 91% and 81% of the patients. The primary end-point was a change in the mean segment diameter of non-dilated segments. The mean lovastatin dose was 33 mg. day(-1). Total- and LDL-cholesterol decreased by 21% and 29% in the special care group and by 7% and 11% in the usual care patients. After 2 years, the mean segment diameter of non-dilated segments decreased by 0.03 mm in the usual care group and 0.004 mm in the special care group (P=0.27). The decrease in the mean segment diameter of dilated lesions was 0.17 mm (usual care) and 0.06 mm (special care) (P=0.04) after 4 months; 0.16 mm (usual care) and 0. 002 mm (special care) after 24 months, respectively (P=0.05). In both groups, the mean segment diameter of dilated lesions increased between 4 and 24 months after PTCA compared to a decrease in mean segment diameter of non-dilated segments (P<0.05). Restenosis (>50% diameter stenosis at follow-up) occurred in 28.4% of usual care and 22.2% of special care patients (P=0.17). CONCLUSIONS: Lovastatin reduced the progression of dilated lesions in men with elective PTCA. Independent of treatment allocation, the dilated lesions regressed and the non-dilated segments progressed during the study follow-up. Four weeks of pre-treatment with lovastatin did not influence the rate of restenosis. Lovastatin had no statistically significant effect on non-dilated segments.

Angioplasty, Balloon, Coronary↗

PubMed1994

Angular induction as a function of the length and position of segments and gaps.

The perceptual distortions which are manifested in the Poggendorff illusion can be studied with the use of a more restricted set of stimulus elements. Experiments were designed in which angular induction effects between two line elements, known respectively as the test segment and induction segment, were evaluated. In some stimulus configurations the induction 'segment' consisted of a tandem pair of segments. Previous studies had shown that the induction segment will bias operant judgments of collinearity for a test segment, this effect being a function of the relative angle between the two. Six experiments are reported, in which the length and position of segments in relation to the tip of the test segment were varied. It was found that substantial induction is produced by a very short segment, and that this can bias judgment even when its displacement spans more than 10 deg of visual angle. Several aspects of the data suggest that the strength of effect is a log-linear function of segment position. However, the results from displacement of single or tandem segments do not conform to predictions based on length/response summation, and thus do not support a linear-systems approach. Neural substrates for these interactions are given brief attention.

Adult↗

PubMed1998

A conserved mode of head segmentation in arthropods revealed by the expression pattern of Hox genes in a spider.

Chelicerates constitute a basic arthropod group with fossil representatives from as early as the Cambrian period. Embryonic development and the subdivision of the segmented body region into a prosoma and an opisthosoma are very similar in all extant chelicerates. The mode of head segmentation, however, has long been controversial. Although all other arthropod groups show a subdivision of the head region into six segments, the chelicerates are thought to have the first antennal segment missing. To examine this problem on a molecular level, we have compared the expression pattern of Hox genes in the spider Cupiennius salei with the pattern known from insects. Surprisingly, we find that the anterior expression borders of the Hox genes are in the same register and the same relative segmental position as in Drosophila. This contradicts the view that the homologue of the first antennal segment is absent in the spider. Instead, our data suggest that the cheliceral segment is homologous to the first antennal segment and the pedipalpal segment is homologous to the second antennal (or intercalary) segment in arthropods. Our finding implies that chelicerates, myriapods, crustaceans, and insects share a single mode of head segmentation, reinforcing the argument for a monophyletic origin of the arthropods.

Animals↗

PubMed1989

Diversity and organization of human T cell receptor delta variable gene segments.

Previous studies of the human TCR-delta gene identified a single commonly used V delta segment, denoted V delta 1. To better understand the extent of the human TCR-delta V gene repertoire, TCR-delta transcripts and gene rearrangements were examined in a new panel of cloned human TCR-gamma/delta lymphocytes. Through this analysis we identified and determined the structures of two new V delta segments, denoted V delta 2 and V delta 3. These V delta segments are different from previously characterized V alpha segments, supporting the notion that the human V delta and V alpha repertoires are distinct. Examination of V gamma gene segment usage in these cells reveals that the V delta 2 gene segment is used in conjunction with the V gamma 2 gene segment. Blot hybridization indicates that the V delta 2 gene segment lies between V delta 1 and D delta-J delta-C delta, and within 100 kb of the latter. Analysis of genomic clones indicates that the V delta 3 gene segment lies in an inverted orientation, approximately 2 kb 3' of C delta. This implies that rearrangement of V delta 3 to D delta-J delta-C delta occurs by inversion. Together with previous mapping studies, these results indicate that human V delta segments are dispersed, rather than clustered, within the TCR-alpha/delta locus. The analysis of rearrangements in polyclonal thymocyte DNA suggests that there may be a limited number of additional V delta gene segments yet to be characterized.

Amino Acid Sequence↗

PubMed1979

Light-induced dephosphorylation of two proteins in frog rod outer segments: influence of cyclic nucleotides and calcium.

Two minor proteins of frog rod outer segments become phosphorylated when retinas are incubated in the dark with 32Pi. The proteins, designated component I (13,000 daltons) and component II (12,000 daltons), are dephosphorylated when retinas are illuminated. The dephosphorylation is reversible; the two proteins are rephosphorylated when illumination ceases. Each outer segment contains approximately 10(6( molecules of components I and II. These remain associated with both fragmented and intact outer segments but dissociate from the outer segment membranes under hypoosmotic conditions. The extent of the light-induced dephosphorylation increases with higher intensities of illumination and is maximal with continuous illumination which bleaches 5.0 x 10(5) rhodopsin molecules/outer segment per second. Light which bleaches 5.0 x 10(3) rhodopsin molecules/outer segment per second causes approximately half-maximal dephosphorylation. This same intermediate level of illumination causes half-suppression of the light-sensitive permeability mechanism in isolated outer segments (Brodie and Bownds. 1976. J. Gen Physiol. 68:1-11) and also induces a half-maximal decrease in their cyclic GMP content (Woodruff et al. 1977. J. Gen. Physiol. 69:667-679). The phosphorylation of components I and II is enhanced by the addition of cyclic GMP or cyclic AMP to either retinas or isolated rod outer segments maintained in the dark. Several pharmacological agents which influence cyclic GMP levels in outer segments, including calcium, cause similar effects on the phosphorylation of components I and II and outer segment permeability. Although the cyclic nucleotide-stimulated phosphorylation can be observed either in retinas or isolated rod outer segments, the light-induced dephosphorylation is observed only in intact retinas.

Animals↗

PubMed1985

Interpretation of myocardial contraction recorded from local segments.

An attempt to quantify the influence of extra-segmental contraction upon external mechanical work of a segment was made in cat hearts. By placing four ultrasound crystals across a borderline between normal and ischaemic myocardium, the pressure-length loop area of segments with different ischaemic contents were recorded. The relation between a normalised segment work index and the amount of ischaemic tissue within the segment showed a negative correlation. Assuming that the contributions to the segment work index from ischaemic and normal tissue coupled in series within the segment add up by cancelling each other, the extra-segmental influence of adjacent fibre contraction could be established. In six cats the extra-segmental influence which affected a typical segment of 10.0 mm corresponded to 1.5 to 3.7 mm of adjacent myocardial musculature on either side of the segment.

Animals↗

PubMed1992

Regional interaction and its effect on patterns of myocardial segmental shortening and lengthening during different models of asynchronous contraction in the dog.

OBJECTIVE: The aim was to examine the effect of asynchrony and regional myocardial interaction on the pattern of segmental contraction and relaxation. METHODS: Three models of asynchrony were produced. Firstly the left anterior descending artery was abruptly occluded for 60 s. Secondly, the same artery was gradually occluded to produce four degrees of ischaemia based on the severity of the mechanical dysfunction. Finally, asynchrony was created by infusing isoprenaline (0.04 microgram.ml-1) into the left circumflex artery. Twelve anaesthetised beagles, weighing 16-21 kg, were used for the study. RESULTS: The patterns of contraction and relaxation were characterised by analysing the phases of shortening and lengthening, the peak lengthening rate (dL/dt), and the timing from the onset of systole to minimum systolic length. A consistent pattern of shortening and lengthening was evident during all three models of asynchrony. There were reciprocal relations between the extent of isovolumetric shortening in the normal segment and in the abnormal segment, and on occasion between the extent of isovolumetric shortening in the normal segment and the extent of isovolumetric lengthening in the same segment. Normal segments that showed minimal shortening or even some lengthening during isovolumetric systole tended to shorten beyond ejection, while segments that shortened significantly during isovolumetric contraction, lengthened earlier. Despite no change in isovolumetric shortening, segments also shortened after ejection when the opposite segment lengthened in late systole and early diastole. CONCLUSIONS: The pattern of shortening and lengthening depends on the path of contraction or on its entire loading pattern throughout systole. It is also possible that during early isovolumetric systole a segment can either be unloaded or preloaded by an opposing segment.

Animals↗

PubMed1991

Fluorescence quenching studies on the characterization of energy generated at the NADH:quinone oxidoreductase and quinol oxidase segments of marine bacteria.

Generation of membrane potential (inside-positive) and delta pH (inside-acidic) at two kinds of NADH:quinone oxidoreductase segments, the Na(+)-motive segment and another segment, of Vibrio alginolyticus was examined by monitoring the quenching of fluorescence of oxonol V and that of quinacrine, respectively, with inside-out membrane vesicles. Transient generation of membrane potential at the segment occurred when ubiquinone-1 was added in the presence of KCN and NADH. The membrane potential was resistant to a proton conductor, carbonylcyanide m-chlorophenylhydrazone, indicating that the membrane potential was generated specifically at the Na(+)-motive segment. On the other hand, neither membrane potential nor delta pH was generated at another segment. The Na(+)-motive segment did not generate delta pH, indicating that only Na+ is extruded at this segment. Furthermore, generation of membrane potential and delta pH at the NADH:quinone oxidoreductase segment of V. anguillarum was examined by using the fluorescence quenching technique. This segment of the bacterium was also found to generate delta psi by the extrusion of Na+ but not H+. These results revealed that the fluorescence quenching technique is useful for the rapid identification and characterization of the respiratory segment involved in Na+ translocation.

Carbonyl Cyanide m-Chlorophenyl Hydrazone↗

PubMed2006

Growth factor-stimulated mitogen-activated kinase (MAPK) phosphorylation in the rat epididymis is limited by segmental boundaries.

Previous evidence has shown that sperm maturation is the result of successive events that influence sperm cells as they move through different microenvironments from the caput to the cauda epididymis. The physiological basis for the creation and maintenance of specific microenvironments along the epididymis are poorly understood. Anatomically, the epididymis consists of segments or lobules of epididymal tubule separated by connective tissue septa (CTS). The fact that CTS restrict the diffusion of tracer substances between segments and that certain gene expression patterns are segment-specific suggest that segments may represent functional epididymal units. In this report, we have further investigated epididymal segmentation by focusing on the ability of CTS to limit the effect of biologically relevant molecules, in particular epidermal growth factor (EGF), basic fibroblast growth factor (FGF2), and vascular endothelial growth factor A (VEGFA), in Segments 1 and 2 of the rat epididymis. We have demonstrated that these growth factors activate mitogen-activated kinase (MAPK) in both segments studied and that growth factors injected into the interstitial space of these segments in vivo exhibited a stimulatory effect only in the segment into which they were injected, i.e., MAPK activation was not observed in the adjacent segment. This restricting influence of CTS was abrogated by treatment with collagenase. In addition, we demonstrate the expression of selected forms of these growth factors and their receptors in Segments 1 and 2, and identify potential downstream targets. These results suggest that CTS regulate the trophic influences of growth factors and potentially other paracrine molecules, thus creating functionally separate units within the epididymis.

Animals↗

PubMed2005

Does anterior plating of the cervical spine predispose to adjacent segment changes?

STUDY DESIGN: In a human cadaveric model, the effects of plate supplementation on the mechanical behaviors of adjacent segments were investigated. OBJECTIVES: The objective was to determine the effects of anterior cervical fusion and plating on the adjacent segments. SUMMARY OF BACKGROUND DATA: Increases in intradiscal pressure and intervertebral motion at adjacent segments have been reported in the lumbar spine following an instrumented fusion. It is unclear if the same phenomenon presents in the cervical spine. METHODS: Seven human cadaveric cervical spines (C2-T1) were used, and fusion of the C5-C6 segment was chosen for the purpose of this study. Two miniature pressure transducers were implanted within each adjacent disc. Flexion, extension, lateral bending, and torsion loads up to 2.5 Nm were applied to the intact spine, and following each of the two procedures, anterior discectomy and grafted fusion, and anterior plating of the C5-C6 motion segment. RESULTS: At the surgical level, a significant increase in segmental stiffness was observed after plating in all directions. Following the grafted fusion, there were no statistically significant changes at the superior adjacent segment, and there was a 13.7% increase in axial rotation in the inferior adjacent segment. Once anterior plating was applied, slight increase (<12%) over the intact spines was noted in lateral bending in both adjacent segments. However, there was no significant difference between the grafted fusion and anterior plated fusion at either adjacent segment. At both adjacent disc levels, the differences in intradiscal pressures between grafted fusion and plated fusion were less than 30% in all directions, and none of these differences was statistically significant. CONCLUSIONS: Intradiscal pressures and intervertebral motion at the adjacent levels are not significantly affected by the instrumented anterior fusion. The clinically observed degenerative change at adjacent segments in the cervical spine is more likely to be attributed to natural progression of the spondylotic process as opposed to biomechanical effect of the instrumentation or fusion.

Aged↗

PubMed1990

Sequence of the large double-stranded RNA segment of the N1 strain of infectious pancreatic necrosis virus: a comparison with other Birnaviridae.

The cDNA sequence of the large dsRNA segment (segment A) of the N1 strain of infectious pancreatic necrosis virus (IPNV) has been determined. The nucleotide and deduced amino acid sequences were compared to the sequences of segment A of the Jasper strain of IPNV and to the sequences of segments A and B (5' and 3' flanking regions) of the 002-73 strain of infectious bursal disease virus (IBDV). The comparison demonstrated that the precursor protein of the major structural polypeptide, pVP2, is highly conserved at the N and C termini, whereas the amino acid sequence of an internal segment shows greater diversity between the strains. This internal segment probably carries the serotype-specific epitopes of birnaviruses. An alternative open reading frame (ORF) (444 bp) partly overlapping with the large ORF (2916 bp) of segment A was found to be conserved among the IPNV strains and is probably also present in the 002-73 strain of IBDV. This small ORF may encode a novel birnavirus polypeptide with an Mr of 17K. SDS-PAGE of radiolabelled purified IPNV particles revealed a band corresponding to the possible novel 17K polypeptide. Short terminal inverted repeats are found in segment A of the N1 and Jasper strains of IPNV and in segment B of the 002-73 strain of IBDV. Segment A of IPNV and segment B of IBDV also contain adjacent inverted repeats at their 5'-terminal flanking regions.

Amino Acid Sequence↗

PubMed1975

Rapid Initiation of Thymidine Incorporation into Deoxyribonucleic Acid in Vegetative Tobacco Stem Segments Treated with Indole-3-acetic Acid.

The short term effect of 11.4 mum indoleacetic acid on the incorporation of (methyl-(3)H)thymidine into DNA in vegetative tobacco (Nicotiana tabacum cv. Wis. 38) stem segments has been investigated. In segments that are defoliated, inverted, and kept in the dark for 7 hours, indoleacetic acid very rapidly (about 60 minutes) and strikingly initiates thymidine incorporation into DNA. The time required before enough indoleacetic acid (2.8 mum) to enhance thymidine incorporation moves into a segment has been found to be about 35 minutes. The initiation response time for segment tissue that already contains 2.8 mum indoleacetic acid should be no more than about 25 minutes. The rate of labeled thymidine incorporation into DNA is affected by physiological treatments of segments. Moving segments from the light into the dark or defoliating segments or inverting defoliated segments decreases the rate of thymidine incorporation. For segments given all three treatments, indoleacetic acid restores the rate of thymidine incorporation as compared to controls. Darkness, or defoliation or inversion of segments, therefore, may decrease thymidine incorporation into DNA by effecting reduced auxin levels in stem segments.

Journal Article↗

PubMed2003

A genetic segmentation of ECG signals.

This paper is concerned with a development of a segmentation technique for electrocardiogram (ECG) signals. Such segmentation is aimed at a lossy signal compression in which each segment can be captured by a simple geometric construct such as, e.g., a linear or quadratic function. The crux of the proposed construct lies in the determination of the optimal segments of data over which they exhibit the highest possible monotonicity (or lowest variability) of the ECG signal. In this sense, the proposed approach generalizes a fundamental and commonly encountered problem of function (data) linearization. The segments are genetically developed using a standard technique of genetic algorithms (GAs). The two fundamental GA constructs, namely a topology of a chromosome and a fitness function governing the optimization process are discussed in detail. The chromosome being coded as a series of floating point numbers contains the endpoints of the segments (segmentation points). The fitness function to be maximized quantifies a level of monotonicity of the ECG data encountered within the segments and takes into consideration differences between the extreme values (minimum and maximum) of its derivatives. As a result of the genetic optimization, we build segments of ECG signals encompassing monotonic (increasing or decreasing) regions of the signal exhibiting a minimal level of variability. A series of experiments dealing with several classes of ECG signals (namely, normal, left bundle branch block beat, and right bundle branch block beat) visualize the effectiveness of the approach and shows the specificity of the linear segments of data. Furthermore, we elaborate on the relationship between the values of the fitness function and the approximation capabilities (quantified by a sum of squared errors between the local model and the data) of the segments of the signal and show that these two descriptors are highly related.

Algorithms↗

PubMed2005

Toward automated segmentation of the pathological lung in CT.

Conventional methods of lung segmentation rely on a large gray value contrast between lung fields and surrounding tissues. These methods fail on scans with lungs that contain dense pathologies, and such scans occur frequently in clinical practice. We propose a segmentation-by-registration scheme in which a scan with normal lungs is elastically registered to a scan containing pathology. When the resulting transformation is applied to a mask of the normal lungs, a segmentation is found for the pathological lungs. As a mask of the normal lungs, a probabilistic segmentation built up out of the segmentations of 15 registered normal scans is used. To refine the segmentation, voxel classification is applied to a certain volume around the borders of the transformed probabilistic mask. Performance of this scheme is compared to that of three other algorithms: a conventional, a user-interactive and a voxel classification method. The algorithms are tested on 10 three-dimensional thin-slice computed tomography volumes containing high-density pathology. The resulting segmentations are evaluated by comparing them to manual segmentations in terms of volumetric overlap and border positioning measures. The conventional and user-interactive methods that start off with thresholding techniques fail to segment the pathologies and are outperformed by both voxel classification and the refined segmentation-by-registration. The refined registration scheme enjoys the additional benefit that it does not require pathological (hand-segmented) training data.

Algorithms↗