Contour Distance Research Articles

A contour distance-based approach for multi-oriented and multi-sized character recognition

In this paper, we propose a novel scheme towards the recognition of multi-oriented and multi-sized isolated characters of printed script. For recognition, at first, distances of the outer contour points from the centroid of the individual characters are calculated and these contour distances are then arranged in a particular order to get size and rotation invariant feature. Next, based on the arranged contour distances, the features are derived from different class of characters. Finally, we use these derived features of the characters to statistically compare the features of the input character for recognition. We have tested our scheme on printed Bangla and Devnagari multi-oriented characters and we obtained encouraging results.

LV Challenge LKEB Contribution: Fully Automated Myocardial Contour Detection

In this paper a contour detection method is described and evaluated on the evaluation data sets of the Cardiac MR Left Ventricle Segmentation Challenge as part of MICCAI 2009’s 3D Segmentation Challenge for Clinical Applications. The proposed method, using 2D AAM and 3D ASM, performs a fully automated detection of the myocardial contours, not requiring any user interaction. The algorithm’s performance is reported using the metrics provided by the LV Challenge organization. Endocardial contour detection was classified as successful in 86% of the images and epicardial contours in 94%. The average perpendicular distance (APD) of the successful contours was 2.28 mm and 2.29 mm for the endo- and epicardial contours, respectively.

Collapse of Linear Polyelectrolyte Chains in a Poor Solvent: When Does a Collapsing Polyelectrolyte Collect its Counterions?

To better understand the collapse of polyions in poor solvent conditions the effective charge and the solvent quality of the hypothetically uncharged polymer backbone need to be known. In the present work this is achieved by utilizing poly-2-vinylpyridine quaternized to 4.3% with ethylbromide. Conductivity and light scattering measurements were utilized to study the polyion collapse in isorefractive solvent/nonsolvent mixtures consisting of 1-propanol and 2-pentanone, respectively, at nearly constant dielectric constant. The solvent quality of the uncharged polyion could be quantified which, for the first time, allowed the experimental investigation of the effect of the electrostatic interaction prior and during polyion collapse, by comparing to a newly developed theory. Although the Manning parameter for the investigated system is as low as lB/l = 0.6 (lB is the Bjerrum length and l is the mean contour distance between two charges), that is, no counterion binding should occur, a qualitative interpretatio...

Denoising and 4D visualization of OCT images

We are using Optical Coherence Tomography (OCT) to image structure and function of the developing embryonic heart in avian models. Fast OCT imaging produces very large 3D (2D + time) and 4D (3D volumes + time) data sets, which greatly challenge ones ability to visualize results. Noise in OCT images poses additional challenges. We created an algorithm with a quick, data set specific optimization for reduction of both shot and speckle noise and applied it to 3D visualization and image segmentation in OCT. When compared to baseline algorithms (median, Wiener, orthogonal wavelet, basic non-orthogonal wavelet), a panel of experts judged the new algorithm to give much improved volume renderings concerning both noise and 3D visualization. Specifically, the algorithm provided a better visualization of the myocardial and endocardial surfaces, and the interaction of the embryonic heart tube with surrounding tissue. Quantitative evaluation using an image quality figure of merit also indicated superiority of th new algorithm. Noise reduction aided semi-automatic 2D image segmentation, as quantitatively evaluated using a contour distance measure with respect to an expert segmented contour. In conclusion, the noise reduction algorithm should be quite useful for visualization and quantitative measurements (e.g., heart volume, stroke volume, contraction velocity, etc.) in OCT embryo images. With its semi-automatic, data set specific optimization, we believe that the algorithm can be applied to OCT images from other applications.

SU‐GG‐I‐89: Automatic Segmentation of Risk Structures

Purpose: The clinical implementation of conformal radiotherapy IMRT has substantially increased the need for automated contouring tools. Implementation of image‐guided adaptive RT will further increase the need. The objective of this work is to develop and assess a fully automated method for delineating risk structures on the basis of statistical models and anatomical atlases. Method and Materials: Anatomical variability is modeled from a set of training images from multiple patients, where structures of interest were manually contoured by clinical experts, and salient anatomical landmarks were defined, e.g. the mastoid process. We use landmark correspondences and thin‐plate spline (TPS) approximation to map all training data into a common geometrical reference frame and generate a mean volumetric image and mean 3D surfaces; spatial variation and dependency of the landmarks is encoded in a statistical model and their local appearance is derived from the mean image. This model is then used to automatically find the positions of the landmarks in a new patient image. TPS is used to map the contours from the mean volume onto the new volume. Finally, the mapped contours are refined using a model‐based segmentation technique. Results: We evaluated our method for three anatomical regions: head‐neck, pelvis and thorax, on multiple clinical datasets accompanied with manually created contours. Quantitative results where collected for spatial accuracy, both for detection of the landmarks as well as for the final contours. For initialization the distance to ground‐truth was 8.9 ± 4.8 mm, contour distance was 3.5 ± 4.5 mm across all anatomies. Computation time was below 50 seconds in all cases. Conclusion: A fully automated segmentation method was developed and quantitatively evaluated for risk structures from CT images in the head‐neck, pelvis and thorax. Future work will increase consistency and efficiency for structures with poor image contrast.

Human DNA mismatch repair: coupling of mismatch recognition to strand-specific excision.

Eukaryotic mismatch-repair (MMR) proteins MutSα and MutLα couple recognition of base mismatches to strand-specific excision, initiated in vivo at growing 3′ ends and 5′ Okazaki-fragment ends or, in human nuclear extracts, at nicks in exogenous circular substrates. We addressed five biochemical questions relevant to coupling models. Excision remained fully efficient at DNA:MutSα ratios of nearly 1 to 1 at various mismatch-nick distances, suggesting a requirement for only one MutSα molecule per substrate. As the mismatch-nick DNA contour distance D in exogenous substrates increased from 0.26 to 0.98 kbp, initiation of excision in extracts decreased as D−0.43 rather than the D−1 to D−2 predicted by some translocation or diffusion models. Virtually all excision was along the shorter (3′–5′) nick-mismatch, even when the other (5′–3′) path was less than twice as long. These observations argue against stochastically directed translocating/diffusing recognition complexes. The failure of mismatched DNA in trans to provoke excision of separate nicked homoduplexes argues against one-stage (concerted) triggering of excision initiation by recognition complexes acting through space. However, proteins associated with gapped DNA did appear to compete in trans with those in cis to mismatch-associated proteins. Thus, as in Escherichia coli, eukaryotic MMR may involve distinct initial-activation and excision-path-commitment stages.

Some remarks concerning the shape of the source contour with application of the method of fundamental solutions to elastic torsion of prismatic rods

This paper deals with numerical experiments related with the shape of the source contour in the application of the method of fundamental solutions to the elastic torsion of prismatic rods. The following five boundary-value problems (BVPs) connected with torsion are studied: L-section, [-section, +-section, ▪-section and I-section. For all five BVPs examined, the region of cross-section of rods is concave. Both the local and global errors are examined for two basic shapes of the source contour. In the first case, the source contour is a circle and in the second case the source contour is geometrically similar to the boundary contour of the region under consideration. Furthermore, the optimal radius of the source contour, in the case of the circle, or the optimal distance of the source contour from the boundary in the case it is geometrically similar, are studied. An influence of the method parameters (radius of the circle or distance between contours) on the condition linear system of equation is examined. In all examples examined the values of the local and global errors of the method are smaller when the source contour is geometrically similar to the boundary of the region under consideration in comparison to the source contour with a shape of a circle.

82: Evaluation of a Semi-Automated Segmentation Method for Delineation of Organs at Risk and Lymph Node Target Volumes in Head and Neck Radiotherapy Planning

The clinical implementation of conformal radiotherapy and IMRT has substantially increased the demand for automated contouring tools. The contouring for head and neck (HN) treatment planning requires many hours of meticulous manual delineation on individual slices of CT data. The objective of this work is to develop and assess an automated method to efficiently and accurately delineate organs-at-risk and uninvolved (N0) lymph node regions in the head and neck. A model-based segmentation framework was developed that deforms a standard HN model to conform to the corresponding structures in individual CT data using a minimum of manual segmentation. Planning CT data of seven patients with small volume primary disease (T1-2) in the nasopharynx, oropharynx, or glottis and N0 levels I-V were retrospectively selected. These patients had normal organs-at-risk and preserved N0 lymph node anatomy to limit the degree of geometrical variability for an initial segmentation study. Manual contouring was carried out for external contour, brain, brainstem, spinal cord, parotids, mandible, clavicles, and N0 lymph node regions Ib-V (International Consensus Guidelines) for all patients. The Ib-V nodal contours were combined to form single right and left volumes. In addition, for model generation, the anatomically bounding structures of the nodal regions such as submandibular gland and hyoid were contoured in two cases. The model stored a multi-modal set of prior knowledge: typical shape of each structure represented by surface meshes, geometrical orientation of structures with respect to one another, and object-specific CT contrast characteristics as a function of location on the mesh surfaces. The model generated from the first case was applied to the remaining cases. The 2nd case was used for model generation to assess the accuracy of the segmentation method for the first case. After manual translation and rotation of the model to the CT anatomy of each case, the algorithm automatically deformed the model to adapt to corresponding boundaries in the image. The algorithm only required manual segmentation of the parotids for anatomical guidance. The segmentation accuracy of the node regions was assessed by compiling the distances between each point on the surface of the auto-segmentation and the closest point on the manual contours. The duration of the procedure was recorded. The average discrepancy between automated and manual segmentation was .24 ± .28 cm (SD) for left Ib-V, and .21 ± .23 cm (SD) for right Ib-V. The maximum error ranged from 1.9 cm (left level V) to 0.7 cm (right level III). On average, 75% (left) and 84% (right) of the nodal area surfaces were within .3 cm distance of the manual contours. The entire segmentation procedure (positioning and adaptation) was performed on average within 118 seconds ± 46 seconds (SD). A model-based segmentation method was developed and quantitatively evaluated for lymph node regions from CT images in the head and neck. Future work will increase the degree of automation, and expand the applicability of the model to include patients with normal variations in anatomy, accommodation of primary disease, and variable degree of nodal involvement as well as consistency and efficiency of output.

A genetic algorithm for the inter-cell layout and material handling system design

The traditional optimization processes for the cell system layout (CSL) and material handling system in cellular manufacturing systems (CMSs) were carried out sequentially and separately. The solutions obtained by this means can be far from the total optimum. In this paper, an integrated approach to the problem is proposed that attempts to design CSL and flow path structure simultaneously. The cells in question are assumed to be shape-fixed and have pre-determined pickup/delivery (P/D) stations. A sequence-pair based CSL-generating algorithm allows cells to be placed at any possible position on a continuous floor plan and leads to a corresponding feasible CSL. A grid graph for the CSL is accordingly constructed to provide possible flow paths for the layout. Then, the sum of the traveling distances, determined by the shortest path algorithm, is used to evaluate the CSL. These steps are embedded into a genetic algorithm (GA) that searches the solution space to obtain optimal layouts based on the contour distances between the P/D stations. Computational results demonstrate that the run time required to solve the test problems is quite acceptable given the long-term nature of facility layout decisions. Moreover, a comparison of the computational results with the existing methods indicates that the proposed approach is a viable alternative for effectively generating layout designs for CMS.

Shape-based hand recognition

The problem of person recognition and verification based on their hand images has been addressed. The system is based on the images of the right hands of the subjects, captured by a flatbed scanner in an unconstrained pose at 45 dpi. In a pre-processing stage of the algorithm, the silhouettes of hand images are registered to a fixed pose, which involves both rotation and translation of the hand and, separately, of the individual fingers. Two feature sets have been comparatively assessed, Hausdorff distance of the hand contours and independent component features of the hand silhouette images. Both the classification and the verification performances are found to be very satisfactory as it was shown that, at least for groups of about five hundred subjects, hand-based recognition is a viable secure access control scheme.

Réduction de speckle et modélisation pour la segmentation d'images échographiques de la prostate

This article discusses a method for automatic prostate segmentation from abdominal ultrasound images. A statistical model of prostate is estimated from training set of manually delineated images. The segmentation starts by smoothing the image to enhance edges by applying a special filter designed for abdominal ultrasound images of prostate. It detects individual speckles and removes them, while it preserves valuable details. Then the boundary is initialised starting from the model and an heuristic optimisation scheme is applied to guide the search of the final form. The performances of the algorithm were compared with manual segmentation, the average contour distance was 2.0 mm and the coverage index was about 93%.

An integrated approach for the concurrent determination of the block layout and the input and output point locations based on the contour distance

This paper presents an integrated approach for the facilities design problem. It develops a method for the concurrent determination of the block layout, the locations of departmental input and output (I/O) points using the contour distances between the I/O points, and the material flow paths between the I/O points. The topology of block layouts is represented using two linear sequences (sequence-pair), which allows the layout to have either a slicing or a non-slicing structure. The block layout is obtained from the sequence-pair with a linear programming formulation. Three heuristic methods are then presented to determine for a given block layout the locations of the I/O points on the perimeters of the departments. The flow paths from output to input points are found by determining the shortest paths that follow the perimeters of the departments. The linear programming algorithm, the shortest path algorithm, and the I/O point location heuristics are embedded into a simulated annealing algorithm that modifies the sequence-pair to obtain a high-quality layout based on the contour distances between the I/O points. Results of computational experiments show that the performance of this integrated algorithm compares favourably with those of algorithms using a sequential approach and is capable of solving industrial-sized problems in acceptable computation time.

Study of Pair Contact Formation among Hydrophobic Residues in a Model HP-36 Protein: Relationship between Contact Order Parameter and Rate of Folding and Collapse

Recent analysis of the structures of a large number of proteins in their native state has demonstrated a close relation between relative contact order parameter (COP) (which gives the average contact distance among hydrophobic residues) and the rate of protein folding (Grantcharova et al. Curr. Opin. Struct. Biol. 2001, 11, 70). We have explored the existence of such a relationship by carrying out Brownian dynamics simulations of a model protein. The model consists of 36 amino acid residues and mimics a thermostable single domain headpiece of chicken villin (HP-36) protein. Long range interactions across the protein are obtained by using a simplified hydropathy scale. The heteropolymer exhibits qualitative features of folding and correlates with the COP. We have defined and calculated the distance dependent pair correlation function among the hydrophobic residues. Nativelike states are characterized by distinct pair contacts which are mostly absent in other collapsed non-native states. Contact pair formation shows a strong dependence on contour distance separation between the pairs. Study of the dynamics of specific contact pair formation during folding shows different characteristics depending on whether the final state is (near) native or far from the native configuration. Approach to the final collapsed state is almost always faster when the final state is nativelike. Fluctuations in pair contacts are found to be smaller in nativelike conformations compared to those of higher energy configurations. This is in agreement with a recent theoretical prediction.

A new method for representing and matching shapes of natural objects

A new method for the representation and comparison of irregular two-dimensional shapes is presented. This method uses a polar transformation of the contour points about the geometric centre of the object. The distinctive vertices of the shape are extracted and used as comparative parameters to minimize the difference of contour distance from the centre. Experiments are performed, more than 39 000 comparisons of database shapes, provided by Sebastian et al. (ICCV (2001) 755), are made and the results are compared to those obtained therein. In addition, 450 comparisons of leaf shape are made and leaves of very similar shape are accurately distinguished. The method is shown to be invariant to translation, rotation and scaling and highly accurate in shape distinction. The method shows more tolerance to scale variation than that of Sebastian et al. (ICCV (2001) 755) and is less computationally intense.

Shape based leaf image retrieval

The authors present an efficient two-stage approach for leaf image retrieval by using simple shape features including centroid–contour distance (CCD) curve, eccentricity and angle code histogram (ACH). In the first stage, the images that are dissimilar with the query image will be first filtered out by using eccentricity to reduce the search space, and fine retrieval will follow by using all three sets of features in the reduced search space in the second stage. Different from eccentricity and ACH, the CCD curve is neither scaling-invariant nor rotation-invariant. Therefore, normalisation is required for the CCD curve to achieve scaling invariance, and starting point location is required to achieve rotation invariance with the similarity measure of CCD curves. A thinning-based method is proposed to locate starting points of leaf image contours, so that the approach used is more computationally efficient. Actually, the method can benefit other shape representations that are sensitive to starting points by reducing the matching time in image recognition and retrieval. Experimental results on 1400 leaf images from 140 plants show that the proposed approach can achieve a better retrieval performance than both the curvature scale space (CSS) method and the modified Fourier descriptor (MFD) method. In addition, the two-stage approach can achieve a performance comparable to an exhaustive search, but with a much reduced computational complexity.

Hand tracking for behaviour understanding

A real-time computer vision system is described for tracking hands thus enabling behavioural events to be interpreted. Forearms are tracked to provide structural context, enabling mutual occlusion, which occurs when hands cross one another, to be handled robustly. No prior skin colour models are used. Instead adaptive appearance models are learned on-line. A contour distance transform is used to control model adaptation and to fit 2D geometric models robustly. Hands can be tracked whether clothed or unclothed. Results are given for a ‘smart desk’ and an in-vehicle application. The ability to interpret behavioural events of interest when tracking a vehicle driver's hands is described.

Surgical correction of the face with the square jaw and weak chin: angle-to-chin bone transfer.

A square-jawed face with a prominent mandibular angle gives a strong, masculine impression. In women, this is an aesthetic problem that is often due to a protruding hyperostotic bone at the angle of the mandible, which can be successfully treated with a mandibular angle osteotomy and bony contouring through an intraoral approach. When square jaws are combined with a weak or deficient chin, the facial balance becomes worse, accentuating the vertical and horizontal lower facial disproportion. To correct these combined problems at the same time, the protruding mandibular angle was resected by a curvilinear or tangential osteotomy, and the bone from the mandibular angle osteotomy was used as an interpositional bone graft to augment a vertically or horizontally deficient chin after a horizontal osteotomy of the chin. From 1991 to 1998, 33 consecutive patients with a prominent mandibular angle and short chin underwent this operation. The square appearance of the lower face was converted to a more graceful, ovoid, and thinner contour. The mean amount of horizontal and vertical chin movement was 4.3 mm and 7.0 mm, respectively. The aesthetic results, as determined by both patient and surgeon, were satisfactory in 30 of 33 cases during the follow-up period of 6 month to 6 years, with minimal complications. The advantages of this technique include reduction of wide bigonial distance and smoothening of square lower facial contour; vertical and horizontal elongation of the chin and the lower face, which enables visual compensation for sagittal weakness of the lower face; improvement of overall facial balance with a single operative session; no additional bone graft donor-site and morbidity. (Plast. Reconstr. Surg. 108: 225, 2001.)

Integrated facilities design using a contour distance metric

The unequal area facility design problem has been studied in the literature generally by considering block layout separate from detailed design. This paper shows a tractable method for concurrently optimizing department shape and location within the facility, and selection of the appropriate number of Input/Output (I/O) stations per department and then optimally locating them. The distance metric used is the contour distance that measures transport between I/O stations along the perimeters of departmental boundaries. This distance metric better reflects allowable material flow for many industrial applications and it directly implies the optimal aisle structure for the facility.

Cardiac and lung complication probabilities after breast cancer irradiation

Purpose: To assess for locoregional irradiation of breast cancer patients, the dependence of cardiac (cardiac mortality) and lung (radiation pneumonitis) complications on treatment technique and individual patient anatomy.Materials and methods: Three-dimensional treatment planning was performed for 30 patients with left-sided breast cancer and various breast sizes. Two locoregional techniques (Techniques A and B) and a tangential field technique, including only the breast in the target volume, were planned and evaluated for each patient. In both locoregional techniques tangential photon fields were used to irradiate the breast. The internal mammary (IM)–medial supraclavicular (MS) lymph nodes were treated with an anterior mixed electron/photon field (Technique A) or with an obliquely incident mixed electron/photon IM field and an anterior electron/photon MS field (Technique B). The optimal IM and MS electron field dimensions and energies were chosen on the basis of the IM–MS lymph node target volume as delineated on CT-slices. The position of the tangential fields was adapted to match the IM–MS fields. Dose-volume histograms (DVHs) and normal tissue complication probabilities (NTCPs) for the heart and lung were compared for the three techniques. In the beam's eye view of the medial tangential fields the maximum distance of the heart contour to the posterior field border was measured; this value was scored as the Maximum Heart Distance.Results: The lymph node target volume receiving more than 85% of the prescribed dose was on average 99% for both locoregional irradiation techniques. The breast PTV receiving more than 95% of the prescribed dose was generally smaller using Technique A (mean: 90%, range: 69–99%) than using Technique B (mean: 98%, range: 82–100%) or for the tangential field technique (mean: 98%, range: 91–100%). NTCP values for excess cardiac mortality due to acute myocardial ischemia varied considerably between patients, with minimum and maximum values of 0.1 and 7.5% (Technique A), 0.1 and 5.8% (Technique B) and 0.0 and 6.1% (tangential tech.). The NTCP values were on average significantly higher (P<0.001) by 1.7% (Technique A) and 1.0% (Technique B) when locoregional breast irradiation was given, compared with irradiation of the left breast only. The NTCP values for the tangential field technique could be estimated using the Maximum Heart Distance. NTCP values for radiation pneumonitis were very low for all techniques; between 0.0 and 1.0%.Conclusions: Technique B results in a good coverage of the breast and locoregional lymph nodes, while Technique A sometimes results in an underdosage of part of the target volume. Both techniques result in a higher probability of heart complications compared with tangential irradiation of the breast only. Irradiation toxicity for the lung is low in all techniques. The Maximum Heart Distance is a simple and useful parameter to estimate the NTCP values for cardiac mortality for tangential breast irradiation.

Contour matching using epipolar geometry

Matching features computed in images is an important process in multiview image analysis. When the motion between two images is large, the matching problem becomes very difficult. In this paper, we propose a contour matching algorithm based on geometric constraints. With the assumption that the contours are obtained from images taken from a moving camera with static scenes, we apply the epipolar constraint between two sets of contours and compute the corresponding points on the contours. From the initial epipolar constraints obtained from corner point matching, candidate contours are selected according to the epipolar geometry, contour end point constraints, and contour distance measures. In order to reduce the possibility of false matches, the number of match points on a contour is also used as a selection measure. The initial epipolar constraint is refined from the matched sets of contours. The algorithm can be applied to a pair or two pairs of images. All of the processes are fully automatic and successfully implemented and tested with various real images.