The Experts below are selected from a list of 570 Experts worldwide ranked by ideXlab platform

Roberto Cipolla - One of the best experts on this subject based on the ideXlab platform.

  • Multiscale Categorical Object Recognition Using Contour Fragments
    IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008
    Co-Authors: Jamie Shotton, Andrew Blake, Roberto Cipolla
    Abstract:

    Psychophysical studies show that we can recognize objects using fragments of outline contour alone. This paper proposes a new automatic visual recognition system based only on local contour features, capable of localizing objects in space and scale. The system first builds a class-specific codebook of local fragments of contour using a novel formulation of Chamfer Matching. These local fragments allow recognition that is robust to within-class variation, pose changes, and articulation. Boosting combines these fragments into a cascaded sliding-window classifier, and mean shift is used to select strong responses as a final set of detection. We show how learning can be performed iteratively on both training and test sets to bootstrap an improved classifier. We compare with other methods based on contour and local descriptors in our detailed evaluation over 17 challenging categories and obtain highly competitive results. The results confirm that contour is indeed a powerful cue for multiscale and multiclass visual object recognition.

  • shape context and Chamfer Matching in cluttered scenes
    Computer Vision and Pattern Recognition, 2003
    Co-Authors: Arasanathan Thayananthan, Bjorn Stenger, Philip H S Torr, Roberto Cipolla
    Abstract:

    This paper compares two methods for object localization from contours: shape context and Chamfer Matching of templates. In the light of our experiments, we suggest improvements to the shape context: shape contexts are used to find corresponding features between model and image. In real images it is shown that the shape context is highly influenced by clutters; furthermore, even when the object is correctly localized, the feature correspondence may be poor. We show that the robustness of shape Matching can be increased by including a figural continuity constraint. The combined shape and continuity cost is minimized using the Viterbi algorithm on features, resulting in improved localization and correspondence. Our algorithm can be generally applied to any feature based shape Matching method. Chamfer Matching correlates model templates with the distance transform of the edge image. This can be done efficiently using a coarse-to-fine search over the transformation parameters. The method is robust in clutter, however, multiple templates are needed to handle scale, rotation and shape variation. We compare both methods for locating hand shapes in cluttered images, and applied to word recognition in EZ-Gimpy images.

Hanne M Kooy - One of the best experts on this subject based on the ideXlab platform.

  • quantification and predictors of prostate position variability in 50 patients evaluated with multiple ct scans during conformal radiotherapy
    Radiotherapy and Oncology, 1999
    Co-Authors: Michael J Zelefsky, Diane Crean, Gig S Mageras, Olga Lyass, Laura Happersett, Clifton C Ling, Steven A Leibel, Zvi Fuks, Sarah Bull, Hanne M Kooy
    Abstract:

    Abstract Purpose : To determine the extent and predictors for prostatic motion in a large number of patients evaluated with multiple CT scans during radiotherapy, and evaluate the implications of these data on the design of appropriate treatment margins for patients receiving high-dose three-dimensional conformal radiotherapy. Materials and methods : Fifty patients underwent four serial computerized tomography (CT) scans, consisting of an initial planning scan and subsequent scans at the beginning, middle, and end of the treatment course. Each scan was performed with the patient in the prone treatment position within an immobilization device used during therapy. Contours of the prostate and seminal vesicles were drawn on the axial CT slices of each scan, and the scans were matched by alignment of the pelvic bones with a Chamfer Matching algorithm. Using the contour information, distributions of the displacement of the organ center of mass and organ border from the planning position were determined separately for the prostate and seminal vesicles in each of the three principle directions: anterior–posterior (AP), superior–inferior (SI) and left–right (LR). Each distribution was fitted to a normal (Gaussian) distribution to determine confidence limits in the center of mass and border displacements and thereby evaluate for the optimal margins needed to contain target motion. Results : The most common directions of displacement of the prostate center of mass (COM) were in the AP and SI directions and were significantly larger than any LR movement. The mean prostate COM displacement (±1 standard deviation, SD) for the entire population was −1.2±2.9 mm, −0.5±3.3 mm and −0.6±0.8 mm in the, AP and SI and LR directions respectively (negative values indicate posterior, inferior or left displacement). The mean (±1 SD) seminal vesicle COM displacement for the entire population was −1.4±4.9 mm, 1.3±5.5 mm and −0.8±3.1 mm in the AP and SI and LR directions, respectively. The data indicate a tendency for the population towards posterior displacements of the prostate from the planning position and both posterior and superior displacements of the seminal vesicles. AP movement of both the prostate and seminal vesicles were correlated with changes in rectal volume ( P =0.0014 and P =0.030 for seminal vesicles and 0.19 for prostate). A logistic regression analysis identified the combination of rectal volume > 60 cm 3 and bladder volumes > 40 cm 3 as the only predictor of large (> 3 mm) systematic deviations for the prostate and seminal vesicles ( P =0.05) defined for each patient as the difference between organ position in the planning scan and mean position as calculated from the three subsequent scans. Conclusions : Prostatic displacement during a course of radiotherapy is more pronounced among patients with initial planning scans with large rectal and bladder volumes. Such patients may require more generous margins around the CTV to assure its enclosure within the prescription dose region. Identification and correction of patients with large systematic errors will minimize the extent of the margin required and decrease the volume of normal tissue exposed to higher radiation doses.

  • automatic three dimensional correlation of ct ct ct mri and ct spect using Chamfer Matching
    Medical Physics, 1994
    Co-Authors: Marcel Van Herk, Hanne M Kooy
    Abstract:

    Image correlation is often required to utilize the complementary information in CT, MRI, and SPECT. A practical method for automatic image correlation in three-dimensions (3D) based on Chamfer Matching is described. The method starts with automatic extraction of contour points in one modality and automatic segmentation of the corresponding feature in the other modality. A distance transform is applied to the segmented volume and a cost function is defined that operates between the contour points and the distance transform. Matching is performed by iteratively optimizing the cost function for 3D translation, rotation, and scaling of the contour points. The complete Matching process including segmentation requires no user interaction and takes about 100 s on an HP715/50 workstation. Perturbation tests on clinical data with cost functions based on mean, rms, and maximum distances in combination with two general purpose optimization procedures have been performed. The performance of the methods has been quantified in terms of accuracy, capture range, and reliability. The best results on clinical data are obtained with the cost function based on the mean distance and the simplex optimization method. The accuracy is 0.3 mm for CT-CT, 1.0 mm for CT-MRI, and 0.7 mm for CT-SPECT correlation of the head. The accuracy is usually at subpixel level but is limited by global geometric distortions, e.g., for CT-MRI correlation. Both for CT-CT and CT-MRI correlation the capture range is about 6 cm, which is higher than normal differences in patient setup found on the scanners (less than 4 cm). This means that the correlation procedure seldom fails (better than 98% reliability) and user interaction is unnecessary. For CT-SPECT Matching the capture range is about 3 cm (80% reliability), and must be further improved. The method has already been introduced in clinical practice.

  • image fusion for stereotactic radiotherapy and radiosurgery treatment planning
    International Journal of Radiation Oncology Biology Physics, 1994
    Co-Authors: Hanne M Kooy, Marcel Van Herk, Patrick D Barnes, Eben Alexander, Susan F Dunbar, Nancy J Tarbell, Robert V Mulkern, Edward J Holupka, Jay S Loeffler
    Abstract:

    Abstract Purpose: We describe an image fusion application that addresses two basic problems that previously limited the use of magnetic resonance imaging (MRI) for geometric localization in stereotactic radiosurgery (SRS) and stereotactic radiotherapy (SRT). The first limitation is imposed by the use of a relocatable, MRI-incompatible, stereotactic frame for stereotactic radiotherapy. The second limitation is an inherent lack of geometric fidelity in current MRI scanners that invalidates the use of MRI for stereotactic localization. Methods and Materials: We recently developed and implemented a novel automated method for fusing computerized tomography (CT) and MRI volumetric image studies. The method is based on a Chamfer Matching algorithm, and provides a quality assurance procedure to verify the accuracy of the fused image set. The image fusion protocol removes the need for stereotactic fixation of the patient for the MRI study. Results: The image fusion protocol significantly improves on the spatial accuracy of the MRI study. We demonstrate the effect of distortion and the effectiveness of the fusion with a phantom study. We present two case studies, an acoustic neurinoma treated with SRS. and a pilocytic astrocytoma treated with SRT. Conclusion: The image fusion protocol significantly improves our logistical management of treating patients with radiosurgery and makes conformal therapy practical for treating patients with SRT. The image fusion protocol demonstrates both the superior diagnostic quality and the poor geometric fidelity of MRI. MRI is a required imaging modality in stereotactic therapy. Image fusion combines the superior MRI diagnostic quality with the superior CT geometric definition, and makes the use of MRI in stereotactic therapy possible and practical.

David Schreiber - One of the best experts on this subject based on the ideXlab platform.

  • gpu accelerated human detection using fast directional Chamfer Matching
    Computer Vision and Pattern Recognition, 2013
    Co-Authors: David Schreiber, Csaba Beleznai, Michael Rauter
    Abstract:

    We present a GPU-accelerated, real-time and practical, pedestrian detection system, which efficiently computes pedestrian-specific shape and motion cues and combines them in a probabilistic manner to infer the location and occlusion status of pedestrians viewed by a stationary camera. The articulated pedestrian shape is approximated by a mean contour template, where template Matching against an incoming image is carried out using line integral based, Fast Directional Chamfer Matching, employing variable scale templates (hybrid CPU-GPU). The motion cue is obtained by employing a compressed non-parametric background model (GPU). Given the probabilistic output from the two cues, the spatial configuration of hypothesized human body locations is obtained by an iterative optimization scheme taking into account the depth ordering and occlusion status of individual hypotheses. The method achieves fast computation times (32 fps) even in complex scenarios with a high pedestrian density. Employed computational schemes are described in detail and the validity of the approach is demonstrated on three PETS2009 datasets depicting increasing pedestrian density.

  • A GPU accelerated Fast Directional Chamfer Matching algorithm and a detailed comparison with a highly optimized CPU implementation
    IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, 2012
    Co-Authors: Michael Rauter, David Schreiber
    Abstract:

    In this work we present an efficient GPU implementation of the Fast Directional Chamfer Matching (FDCM) algorithm [10]. We propose some extensions to the original FDCM algorithm. In particular, we extend the algorithm to handle templates with variable size, to account for perspective effects. To the best of our knowledge, our work is the first to present a full implementation of a shape based Matching algorithm on a GPU. Further contributions of our work consist of implementing a highly optimized CPU version of the algorithm (via multi-threading and SSE2), as well as a thorough comparison between pure GPU, pure CPU, and a hybrid version. The hybrid CPU-GPU version which turns out to be the fastest, achieves run-time of 44 fps on PAL resolution images.

A Touw - One of the best experts on this subject based on the ideXlab platform.

  • the effect of image artifacts organ motion and poor segmentation on the reliability and accuracy of 3d Chamfer Matching
    CVRMed-MRCAS '97 Proceedings of the First Joint Conference on Computer Vision Virtual Reality and Robotics in Medicine and Medial Robotics and Compute, 1997
    Co-Authors: M Van Herk, Kenneth G A Gilhuijs, J C De Munck, A Touw
    Abstract:

    Chamfer Matching was tested on pairs of pelvic CT scans with a number of artificially and natural artifacts (for instance as model for CT-MR Matching) and was found to be extremely robust against outliers (e.g., due to organ motion), missing data, low resolution, and poor segmentation of the images.

  • effect of image artifacts organ motion and poor segmentation on the reliability and accuracy of three dimensional Chamfer Matching
    Computer Aided Surgery, 1997
    Co-Authors: M Van Herk, Kenneth G A Gilhuijs, J C De Munck, A Touw
    Abstract:

    Our objective was to investigate the influence of various image artifacts on three-dimensional Chamfer Matching. A number of artificial and natural artifacts (for instance, as a model for CT-MR Matching) were introduced or suppressed in pairs of pelvic CT scans, and a perturbation study was used to determine reliability and accuracy in a well known ground truth situation. In general, Chamfer Matching is extremely robust against missing data, low resolution, and poor segmentation of the images. In the presence of artifacts, minimization of the average distance outperformed minimization of the root-mean-square distance. Outliers in the scan from which the point list is obtained must be avoided. For example, rotation of the femurs reduces CT-CT registration accuracy by 1-2 mm. The robustness of Chamfer Matching is confirmed by a limited perturbation study of CT-MR registration for the pelvic region. In conclusion, Chamfer Matching is extremely accurate and reliable if outliers are avoided in the scan from which the point list is derived, and the average distance is used as a cost function.

  • automatic registration of ct and mr of the pelvis using Chamfer Matching
    International Conference of the IEEE Engineering in Medicine and Biology Society, 1996
    Co-Authors: M Van Herk, Joos V. Lebesque, J C De Munck, S H Muller, A Touw
    Abstract:

    The aim of this work is to develop and test an automatic image registration method for CT and MR of the pelvis which does not require external landmarks, and works with MR slices in any orientation (e.g., axial, coronal and saggital). By combining Chamfer Matching with automatic segmentation methods, a fast (a few minutes) and accurate (about 2 mm) method for CT-MR registration is obtained. However, failures of the Matching method due to local minima cannot always be avoided. In such cases, manual adjustment of the starting position is sufficient to obtain a good match.

  • variation in volumes dose volume histograms and estimated normal tissue complication probabilities of rectum and bladder during conformal radiotherapy of t3 prostate cancer
    International Journal of Radiation Oncology Biology Physics, 1995
    Co-Authors: Joos V. Lebesque, A Bruce, A Guus P Kroes, A Touw, Tarek Shouman, Marcel Van Herk
    Abstract:

    Purpose : To determine the pattern of changes of rectum and bladder structures during conformal therapy of T3 prostate cancer and the impact of these changes on the accuracy of the dose-volume histograms (DVHs) and normal tissue complication probabilities (NTCPs) of these organs, based on the planning computed tomography (CT) scan only. Methods and Materials : For 11 T3 prostate cancer patients treated with conformal therapy, three repeat CT scans were made in Weeks 2, 4, and 6 of the treatment. The bony anatomy was aligned with the planning CT scan, using three dimensional (3D) Chamfer Matching. The internal and external surfaces of rectum and bladder were contoured in each scan. Three volumes were calculated for each organ : solid organ (including filling), filling, and wall volume. DVHs and NTCPs were calculated for all structures. Results : The solid organ and filling volumes varied considerably between patients and within a patient and they decreased with increasing treatment time. The largest patient variation was seen for patients with large initial filling volumes. The variations of rectum and bladder wall volumes during treatment were 9 and 17% (1 standard deviation (SD)), respectively, with no time trend. The changes of the high dose (>80 and 90% of the prescribed dose) volumes of the rectum in response to rectum filling differences were proportional to the whole rectum volume changes. The variation of the high-dose rectum wall volume was relatively small (14%, 1 SD). As a result, the NTCPs of rectum and rectum wall were the same overall and the variation of the NTCPs during treatment was about 14% (1 SD) and not correlated with rectum filling. The variation of the high-dose bladder volumes (about 14%, 1 SD) was smaller than the variation of the whole bladder volumes (30%, 1 SD). The high-dose bladder wall volume decreased significantly due to wall distention as the bladder filling increased. As a result of this complex pattern, the variation of NTCPs of bladder (85%, 1 SD) and bladder wall (88%, 1 SD) during treatment was large and significantly correlated with bladder filling. Conclusions : The planning CT scan overestimates rectum and bladder filling during treatment. Furthermore, the variation of filling is so large that only the wall structures have relatively constant volumes during treatment. For the rectum wall, the DVHs and NTCPs, as estimated from the initial scan, are representative for the whole treatment, because no correlation was seen between these parameters and organ filling. For the bladder wall, however, such a correlation was present and consequently, the initial bladder wall DVHs and NTCPs can only be representative for the whole treatment, if the bladder filling can be kept reasonably constant during treatment.

  • quantification of organ motion during conformal radiotherapy of the prostate by three dimensional image registration
    International Journal of Radiation Oncology Biology Physics, 1995
    Co-Authors: Marcel Van Herk, A Bruce, A Guus P Kroes, A Touw, Tarek Shouman, Joos V. Lebesque
    Abstract:

    PURPOSE: Knowledge about the mobility of organs relative to the bony anatomy is of great importance when preparing and verifying conformal radiotherapy. The conventional technique for measuring the motion of an organ is to locate landmarks on the organ and the bony anatomy and to compare the distance between these landmarks on subsequent computerized tomography (CT) scans. The first purpose of this study is to investigate the use of a three dimensional (3D) image registration method based on Chamfer Matching for measurement of the location and orientation of the whole organ relative to the bony anatomy. The second purpose is to quantify organ motion during conformal therapy of the prostate. METHODS AND MATERIALS: Four CT scans were made during the course of conformal treatment of 11 patients with prostate cancer. With the use of a 3D treatment planning system, the prostate and seminal vesicles were contoured interactively. In addition, bladder and rectum were contoured and the volume computed. Next, the bony anatomy of subsequent scans was segmented and matched automatically on the first scan. The femora and the pelvic bone were matched separately to quantify motion of the legs. Prostate (and seminal vesicle) contours from the subsequent scans were matched on the corresponding contours of the first scan, resulting in the 3D rotations and translations that describe the motion of the prostate and seminal vesicles relative to the pelvic bone. RESULTS: Bone Matching of two scans with about 50 slices of 256 x 256 pixels takes about 2 min on a workstation and achieves subpixel registration accuracy. Matching of the organ contours takes about 30 s. The accuracy in determining the relative movement of the prostate is 0.5 to 0.9 mm for translations (depending on the axis) and 1 degree for rotations (standard deviations). Because all organ contours are used for Matching, small differences in delineation of the prostate, missing slices, or differences in slice distance have only a limited influence on the accuracy. Rotations of the femora and the pelvic bone are quantified with about 0.4 degree accuracy. A strong correlation was found between rectal volume and anterior-posterior translation and rotation around the left-right axis of the prostate. Consequently, these parameters had the largest standard deviations of 2.7 mm and 4.0 degrees. Bladder filling had much less influence. Less significant correlations were found between various leg rotations and pelvic and prostate motion. Standard deviations of the rotation angles of the pelvic bone were less than 1 degree in all directions. CONCLUSIONS: Using 3D image registration, the motion of organs relative to bony anatomy has been quantified accurately. Uncertainties in contouring and visual interpretation of the scans have a much smaller influence on the measurement of organ displacement with our new method than with conventional methods. We have quantified correlations between rectal filling, leg motions, and prostate motion.

Ioannis Pitas - One of the best experts on this subject based on the ideXlab platform.

  • optimized Chamfer Matching for snake based image contour representations
    International Conference on Multimedia and Expo, 2006
    Co-Authors: Andras Hajdu, A Roubies, Ioannis Pitas
    Abstract:

    In this paper we present a novel method on how to take advantage of the snake representation of target objects, when doing Chamfer Matching for detection/recognition purposes. In this case several time-consuming steps of classic Chamfer Matching approaches can be simplified. Moreover, we investigate the possibility of involving fewer pixels from both the target and template object to speed up computations. We introduce an optimization method for such an object reduction, which is valid also in the general application scheme of Chamfer Matching. Finally, we present our experimental results regarding human body detection