No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Context specific descriptors for tracking deforming tissue
Abstract
In minimally invasive surgery, deployment of motion compensation, dynamic active constraints and adaptive intra-operative guidance require accurate estimation of deforming tissue in 3D. To this end, the use of vision-based techniques is advantageous in that it does not require the integration of additional hardware to the existing surgical settings. Deformation can be recovered by tracking features on the surface of the tissue. Existing methods are mostly based on ad hoc machine vision techniques that have generally been developed for rigid scenes or use pre-determined models with parameters fine tuned to specific surgical settings. In this work, we propose a novel tracking technique based on a context specific feature descriptor. The descriptor can adapt to its surroundings and identify the most discriminate information for tracking. The feature descriptor is represented as a decision tree and the tracking process is formulated as a classification problem for which log likelihood ratios are used to improve classifier training. A recursive tracking algorithm obtains examples of tissue deformation used to train the classifier. Additional training data is generated by geometric and appearance modelling. Experimental results have shown that the proposed context specific descriptor is robust to drift, occlusion, and changes in orientation and scale. The performance of the algorithm is compared with existing tracking algorithms and validated with both simulated and in vivo datasets.
... In addition, depending on the applications (for example 3D depth recovery), it is desirable that the tracked features span over the field of view. Currently, research in the field of feature tracking on MIS images [12][13][14][15][16][17][18] faces many problems essentially due to their image properties. Unlike synthetic scenes and outdoor environments, MIS images exhibit considerable drawbacks: specular reflection, shadow effects, dynamic lighting conditions, non-rigid http://dx.doi.org/10.1016/j.compmedimag.2014. ...
... It is generally sensitive to the pair detector and tracking algorithms. Previous work [15,14,17,13] usually compares the number of detected points using different detectors with their default parameters and reports the percentage of robust ones. Generally, the compared sets do not have the same number of points at the detection step. ...
... Other approaches [19,14,15] evaluate the tracking performance using a 3D rigid model or phantom. They compute the registration error between the scanned 3D model and the surface reconstructed from the tracked points using Structure From Motion or Stereovision. ...
... These conditions may result in tracking failure. Several research studies attempt [13,21,20,16] to identify the best features to track from endoscopic images using optical flow and/or feature descriptor methods. Nevertheless, they have not yet resulted in a routinely used clinical application due to MIS image properties. ...
... Evaluation of the Detection Quality: Previous work [21,13,16] usually compares the number of detected points for different detectors and reports the percentage of the robust ones. No quantitative analysis to measure the distribution of points is provided and it do not investigate the choice of algorithm parameter values. ...
... In addition, manual labelling could be only done on salient features, which is difficult for the liver due to its homogeneous surface texture. Other approaches [8,13,21] evaluate the tracking performance using a 3D rigid model or phantom. They compute the registration error between the reconstructed surface (using Structure From Motion or Stereo) from the tracked points and the scanned 3D model. ...
Feature tracking for endoscopic images is a critical component for image guided applications in minimally invasive surgery. Recent work in this field has shown success in acquiring tissue deformation, but it still faces issues. In particular, it often requires expensive algorithms to filter outliers. In this paper, we firstly propose two real-time pre-processes based on image filtering, to improve feature tracking robustness and thus reduce outlier percentage. However the performance evaluation of detection and tracking algorithms on endoscopic images is still difficult and not standardized, due to the difficulty of ground truth data acquisition. To overcome this drawback, we secondly propose a novel framework that allows to provide artificial ground truth data, and thus to evaluate detection and feature tracking performances. Finally, we demonstrate, using our framework on 9 different in-vivo video sequences, that the proposed pre-processes significantly increase the tracking performance.
... Practical implementations of the state-of-the-art techniques remain non-trivial in the case of close-range scene with varying texture level during endoscopic imaging of the placenta despite the extensive study of feature-based tracking for surgical navigation [18][19][20] including image alignment application [20][21][22]. While many researches [18,19] in this field focus on modeling tissue deformation to enhance scene adaptability, the use of fixed generalized parameters is often susceptible to detection and matching failure due to unpredictable imaging conditions of the scene associated with placental imaging. ...
... Practical implementations of the state-of-the-art techniques remain non-trivial in the case of close-range scene with varying texture level during endoscopic imaging of the placenta despite the extensive study of feature-based tracking for surgical navigation [18][19][20] including image alignment application [20][21][22]. While many researches [18,19] in this field focus on modeling tissue deformation to enhance scene adaptability, the use of fixed generalized parameters is often susceptible to detection and matching failure due to unpredictable imaging conditions of the scene associated with placental imaging. Such images are characterized by large variations in the level of texture and frequent shortage of distinctive correspondence even amongst temporally adjacent views of the fetoscopic trajectory [11,23]. ...
Background:
Visualization of the vast placental vasculature is crucial in fetoscopic laser photocoagulation for twin-to-twin transfusion syndrome treatment. However, vasculature mosaic is challenging due to the fluctuating imaging conditions during fetoscopic surgery.
Method:
A scene adaptive feature-based approach for image correspondence in free-hand endoscopic placental video is proposed. It contributes towards existing techniques by introducing a failure detection method based on statistical attributes of the feature distribution, and an updating mechanism that self-tunes parameters to recover from registration failures.
Results:
Validations on endoscopic image sequences of a phantom and a monkey placenta are carried out to demonstrate mismatch recovery. In two 100-frame sequences, automatic self-tuned results improved by 8% compared with manual experience-based tuning and a slight 2.5% deterioration against exhaustive tuning (gold standard).
Conclusion:
This scene-adaptive image correspondence approach, which is not restricted to a set of generalized parameters, is suitable for applications associated with dynamically changing imaging conditions. Copyright © 2015 John Wiley & Sons, Ltd.
... They have considered tissue tracking as a classification problem and solved it using decision trees. They have argued that the proposed method is robust to drift, occlusion, orientation and scalingvariations (Mountney & Yang, 2012). Since the classifier trains in a supervised manner, the generalization ability for tissue deformation tracking may be reduced and for new deformations that are not seen in the training set, it may fail to recognize them. ...
... Ground truth segmentation of specular reflection has been performed manually. (Lahane et al., 2012) Unsafe action detection 900 image as train set 213 image as test set (positive and negative classes) (Doignon et al., 2005) segmentation of surgical instruments Images from 3 video sequences of 500 color images (Climent & Mars, 2010) Instrument localization 128 images extracted from a real operation video (Sa-Ing et al., 2012) Instrument tracking simulated videos for different situations (Zhao, 2014) 3D visual tracking of laparoscopic instruments a plastic phantom with a test-bench inside (Collins et al., 2011) deformable 3D surface reconstruction Simulating a 3D kidney model comprising 1820 vertices (Mountney & Yang, 2012) tracking deforming tissue simulated and in vivo MIS datasets of the liver, heart and abdomen (Grasso et al., 2009) Laparoscopic Video Summarization 378 images randomly selected from five laparoscopic cholecystectomy videos (Iakovidis et al., 2010) Endoscopic video summarization annotated video frames with ground truth information labeled manually (Oh et al., 2007) Informative/ non-informative frame classification in endoscopic videos 70 frames were selected as a test set from three colonoscopy videos consisting of 35 informative frames and 35 non-informative frames (Wang et al., 2016) Laparoscopic video summarization and key-frame selection a new gastroscopic video dataset has been proposed from 30 volunteers with more than 400k images (Zappella et al., 2013) Surgical gesture classification Previously presented dataset Many researches have tried to address some of the mentioned issues in their work. But, there is still need to propose more robust and stable methods for any of the mentioned applications in laparoscopic image processing. ...
Laparoscopy is a minimally-invasive surgery using a few small incisions on the patient’s body to insert the tools and telescope and conduct the surgical operation. Laparoscopic video processing can be used to extract valuable knowledge and help the surgeons. We discuss the present and possible future role of processing laparoscopic videos. The various applications are categorized for image processing algorithms in laparoscopic surgeries including preprocessing video frames by laparoscopic image enhancement, telescope related applications (telescope position estimation, telescope motion estimation and compensation), surgical instrument related applications (surgical instrument detection and tracking), soft tissue related applications (soft tissue segmentation and deformation tracking) and high level applications such as safe actions in laparoscopic videos, summarization of laparoscopic videos, surgical task recognition and extracting knowledge using fusion techniques. Some different methods have been proposed previously for each of the mentioned applications using image processing.
... To the best of our knowledge, none of these has yet been applied to endoscopic stitching, despite their high potential. An exception is the work of Mountney et al. [107], [110] (2008, 2011), who have presented an online learning scheme for feature descriptors and adapted the method of randomized trees for keypoint recognition by Lepetit et al. [86] (2006) to develop contextspecific descriptors for application in laparoscopic stereoscopy SLAM. They have also presented a comparison of feature descriptors for MIS [104] (2007). ...
... This is true for all stitching and reconstruction methods that include a global optimization step, such as bundle adjustment. The SLAM methods, primarily applied in laparoscopy, are an exception to this -as presented by Mountney and Yang's group [104]- [110] as a stereoscopic approach, as well as by Grasa et al. [58]- [60] for monoscopic views (see section II-D). Bouma et al. [26] have also presented a realtime reconstruction approach for minimally invasive surgery, which incorporates stereoscopic ego-motion computation in real-time. ...
Endoscopic procedures form part of routine clinical practice for minimally invasive examinations and interventions. While they are beneficial for the patient, reducing surgical trauma and making convalescence times shorter, they make orientation and manipulation more challenging for the physician, due to the limited field of view through the endoscope. However, this drawback can be reduced by means of medical image processing and computer vision, using image stitching and surface reconstruction methods to expand the field of view. This review article provides a comprehensive overview of the current state of the art in endoscopic image stitching and surface reconstruction. The literature in the relevant fields of application and algorithmic approaches is surveyed. The technological maturity of the methods and current challenges and trends are analyzed.
... In the modelbased 3D reconstruction method, since the tracking interest point area is extracted in each frame for calculation, the position tracking of feature points can be obtained using a de- To realize the 3D motion tracking of feature points in different frames, it is necessary to calculate the position of the tracking feature points in different frames. In the model-based 3D reconstruction method, since the tracking interest point area is extracted in each frame for calculation, the position tracking of feature points can be obtained using a deformation model [22,23,29]. In feature-based 3D motion tracking, feature point tracking methods can be roughly divided into two types. ...
At present, feature-based 3D reconstruction and tracking technology is widely applied in the medical field. In minimally invasive surgery, the surgeon can achieve three-dimensional reconstruction through the images obtained by the endoscope in the human body, restore the three-dimensional scene of the area to be operated on, and track the motion of the soft tissue surface. This enables doctors to have a clearer understanding of the location depth of the surgical area, greatly reducing the negative impact of 2D image defects and ensuring smooth operation. In this study, firstly, the 3D coordinates of each feature point are calculated by using the parameters of the parallel binocular endoscope and the spatial geometric constraints. At the same time, the discrete feature points are divided into multiple triangles using the Delaunay triangulation method. Then, the 3D coordinates of feature points and the division results of each triangle are combined to complete the 3D surface reconstruction. Combined with the feature matching method based on convolutional neural network, feature tracking is realized by calculating the three-dimensional coordinate changes of the same feature point in different frames. Finally, experiments are carried out on the endoscope image to complete the 3D surface reconstruction and feature tracking.
... Puerto-Souza and Mariottini (2013) introduced a Hierarchical Multi-Affine (HMA) algorithm to map features between two endoscopic images, allowing to recover features that were lost after a complete occlusion or sudden camera motions. Mountney and Yang (2012) exploited online 85 learning and classification using a context specific feature descriptor, in order to increase the robustness against drift and occlusion. Du et al. (2015) used a triangular geometric mesh model to combine features and intensity information to robustly track soft tissue surface deformation. ...
Despite the benefits introduced by robotic systems in abdominal Minimally Invasive Surgery (MIS), major complications can still affect the outcome of the procedure, such as intra-operative bleeding. One of the causes is attributed to accidental damages to arteries or veins by the surgical tools, and some of the possible risk factors are related to the lack of sub-surface visibilty. Assistive tools guiding the surgical gestures to prevent these kind of injuries would represent a relevant step towards safer clinical procedures. However, it is still challenging to develop computer vision systems able to fulfill the main requirements: (i) long term robustness, (ii) adaptation to environment/object variation and (iii) real time processing. The purpose of this paper is to develop computer vision algorithms to robustly track soft tissue areas (Safety Area, SA), defined intra-operatively by the surgeon based on the real-time endoscopic images, or registered from a pre-operative surgical plan. We propose a framework to combine an optical flow algorithm with a tracking-by-detection approach in order to be robust against failures caused by: (i) partial occlusion, (ii) total occlusion, (iii) SA out of the field of view, (iv) deformation, (v) illumination changes, (vi) abrupt camera motion, (vii), blur and (viii) smoke. A Bayesian inference-based approach is used to detect the failure of the tracker, based on online context information. A Model Update Strategy (MUpS) is also proposed to improve the SA re-detection after failures, taking into account the changes of appearance of the SA model due to contact with instruments or image noise. The performance of the algorithm was assessed on two datasets, representing ex-vivo organs and in-vivo surgical scenarios. Results show that the proposed framework, enhanced with MUpS, is capable of maintain high tracking performance for extended periods of time ( ≃ 4 min - containing the aforementioned events) with high precision (0.7) and recall (0.8) values, and with a recovery time after a failure between 1 and 8 frames in the worst case.
This paper can be freely downloaded from this link https://authors.elsevier.com/a/1WMlU4rfPluH2b until February 28th.
... A comparative evaluation of state-of-the-art feature-matching algorithms for endoscopic images has been carried out in [186]. [155,187,188,205,231,268], deforming tissue tracking is a very hard research challenge that still requires a lot of further work. Endoscopic videos feature many domain-induced problems like scarcity of distinctive landmarks because of homogenous surfaces and indistinctive texture that makes it hard to find good points to track. ...
In recent years, digital endoscopy has established as key technology for medical screenings and minimally invasive surgery. Since then, various research communities with manifold backgrounds have picked up on the idea of processing and automatically analyzing the inherently available video signal that is produced by the endoscopic camera. Proposed works mainly include image processing techniques, pattern recognition, machine learning methods and Computer Vision algorithms. While most contributions deal with real-time assistance at procedure time, the post-procedural processing of recorded videos is still in its infancy. Many post-processing problems are based on typical Multimedia methods like indexing, retrieval, summarization and video interaction, but have only been sparsely addressed so far for this domain. The goals of this survey are (1) to introduce this research field to a broader audience in the Multimedia community to stimulate further research, (2) to describe domain-specific characteristics of endoscopic videos that need to be addressed in a pre-processing step, and (3) to systematically bring together the very diverse research results for the first time to provide a broader overview of related research that is currently not perceived as belonging together.
... A pattern is a set of features that quantify the perceptual information of ROI and the background. The features are in fact descriptors, which can be either quantitative such as area, length and texture [4][5][6], or structural descriptors such as Point symmetry groups and Frieze symmetry groups [7][8][9][10][11]. The pattern recognition algorithms can be therefore fall into two broad classes of decision theoretic and structural pattern recognition, respectively. ...
We developed an adaptive learning for segmentation of plant images into plant and non-plant regions. In this study, we used Kohonen's self organizing map (SOM) algorithm for segmentation of plant images using image series of two complexity levels; images taken in a controlled environment of plant facility and also images taken in the field. Nine color features of three color models of normalized Red Green Blue (RGB), Hue Saturation and Intensity (HSI) and L*a*b* made up the feature map. The results showed good performance for the images with less complexity. However, for images with higher complexity where there are more regions with similar color pattern, the method produces some noise.
Optical molecular imaging (OMI) is an imaging technology that uses an optical signal, such as near-infrared light, to detect biological tissue in organisms. Because of its specific and sensitive imaging performance, it is applied in both preclinical research and clinical surgery. However, it requires heavy data analysis and a complex mathematical model of tomographic imaging. In recent years, machine learning (ML)-based artificial intelligence has been used in different fields because of its ability to perform powerful data processing. Its analytical capability for processing complex and large data provides a feasible scheme for the requirement of OMI. In this paper, we review ML-based methods applied in different OMI modalities.
The massive generation of data has raised new research topics, such as methods to store and to retrieve large volumes of information. Some medical image modalities, such as Magnetic Resonance Imaging, generate hundreds images series and many research groups have presented studies to develop intelligent techniques to classify and to retrieve this information. However, these studies are dispersed in several databases, and cataloged by using different terms. In this paper we present an analysis of these studies, through a Systematic Mapping that identifies methods and techniques currently being used in this scenario. In addition, we provide a perspective about the type of scientific literature the researchers have been disclosed their studies as well as impact on the scientific literature in dissemination of this knowledge domain. Some challenges and research opportunities are also highlighted in order to propitiate advances in the area.
In this paper, we proposed a novel algorithm for endoscopic image matching. The algorithm consists of two main components, log-ratio descriptor and probabilistic matching criterion. Log-ratio descriptor is developed by using selected pair of grayscale intensity information that surround the keypoint. The spatial distribution of the pairs follow approximately normal distribution. Then, probabilistic t-test is implemented to produce a distinctive features descriptor. Acceptable probability is calculated based on the probability of t-distribution information. Finally, matching the keypoints is performed by comparing the acceptable probability and nearest neighbor location information. Simulation results show that the proposed algorithm achieves more than 90% matching in various types of tissue surface and movement.
Intraoperative application of tomographic imaging techniques provides a means of visual servoing for objects beneath the surface of organs.
The focus of this survey is on therapeutic and diagnostic medical applications where tomographic imaging is used in visual servoing. To this end, a comprehensive search of the electronic databases was completed for the period 2000-2013.
Existing techniques and products are categorized and studied, based on the imaging modality and their medical applications. This part complements Part I of the survey, which covers visual servoing techniques using endoscopic imaging and direct vision.
The main challenges in using visual servoing based on tomographic images have been identified. 'Supervised automation of medical robotics' is found to be a major trend in this field and ultrasound is the most commonly used tomographic modality for visual servoing. Copyright © 2014 John Wiley & Sons, Ltd.
A 3D colon model is an essential component of a computer-aided diagnosis (CAD) system in colonoscopy to assist surgeons in visualization, and surgical planning and training. This research is thus aimed at developing the ability to construct a 3D colon model from endoscopic videos (or images). This paper summarizes our on-going research in automated model building in colonoscopy. We have developed the mathematical formulations and algorithms for modeling static, localized 3D anatomic structures within a colon that can be rendered from multiple novel view points for close scrutiny and precise dimensioning. This ability is useful for the scenario when a surgeon notices some abnormal tissue growth and wants a close inspection and precise dimensioning. Our modeling system uses only video images and follows a well-established computer-vision paradigm for image-based modeling. We extract prominent features from images and establish their correspondences across multiple images by continuous tracking and discrete matching. We then use these feature correspondences to infer the camera's movement. The camera motion parameters allow us to rectify images into a standard stereo configuration and calculate pixel movements (disparity) in these images. The inferred disparity is then used to recover 3D surface depth. The inferred 3D depth, together with texture information recorded in images, allow us to construct a 3D model with both structure and appearance information that can be rendered from multiple novel view points.
With a novel approach, abdominal surgery can be per-formed through the natural orifices with a following incision in stomach or colon. "Natural Orifice Translumenal Endoscopic Surgery" (NOTES) is assumed to offer significant benefits to patients such as less pain, faster recovery, and better cosmesis than current laparoscopic techniques. The potential advantages can be only achieved through secure and standardized opera-tion methods. Several barriers identified for the clinical practi-cability in flexible intra-abdominal endoscopy can be solved with computer-assisted surgical systems. For some of them ad-ditional 3-D information is useful, for some it is mandatory. At our institute a 3-D endoscope called Multisensor-Time-Of-Flight (MUSTOF) endoscope is in development. In this paper some MUSTOF based solutions for NOTES will be pointed out: The so obtained 3-D information can be registered with preoperative CT or MR data. These enhanced volumes can be used to find the transgastric or transcolonic entry point or to provide better orientation. It also will enable intra-operative collision prevention and an optimized field of off-axis view. With such an progressive endoscopic system, translumenal surgery will be able to be performed in a safe and feasible manner.
Recovering tissue deformation during robotic assisted minimally in- vasive surgery is an important step towards motion compensation and stabiliza- tion. This paper presents a practical strategy for dense 3D depth recovery and temporal motion tracking for deformable surfaces. The method combines image rectification with constrained disparity registration for reliable depth estimation. The accuracy and practical value of the technique is validated with a tissue phantom with known 3D geometry and motion characteristics. It has been shown that the performance of the proposed approach compares favorably against existing methods. Example results of the technique applied to in vivo robotic assisted minimally invasive surgery data are also provided.
We present the design and a prototype implementation of a three-dimensional visualization system to assist with laparoscopic surgical procedures. The system uses 3D visualization, depth extraction from laparoscopic images, and six degree-of-freedom head and laparoscope tracking to display a merged real and synthetic image in the surgeon’s video-see-through head-mounted display. We also introduce a custom design for this display. A digital light projector, a camera, and a conventional laparoscope create a prototype 3D laparoscope that can extract depth and video imagery.
Such a system can restore the physician’s natural point of view and head motion parallax that are used to understand the 3D structure during open surgery. These cues are not available in conventional laparoscopic surgery due to the displacement of the laparoscopic camera from the physician’s viewpoint. The system can also display multiple laparoscopic range imaging data sets to widen the effective field of view of the device. These data sets can be displayed in true 3D and registered to the exterior anatomy of the patient. Much work remains to realize a clinically useful system, notably in the acquisition speed, reconstruction, and registration of the 3D imagery.
We propose an image-based motion tracking algorithm that can be used with stereo endoscopie and microscope systems. The tracking problem is considered to be a time-varying optimization of a parametric function describing the disparity map. This algorithm could be used as part of a virtual stabilization system that can be employed to compensate residual motion of the heart during robot-assisted off-pump coronary artery bypass surgery (CABG). To test the appropriateness of our methods for this application, we processed an image sequence of a beating pig heart obtained by the stereo endoscope used in the da Vinci robotic surgery system. The tracking algorithm was able to detect the beating of the heart itself as well as the respiration of the lungs.
We explore a new approach to shape recognition based on a virtually infinite family of binary features (queries) of the image data, designed to accommodate prior information about shape invariance and regularity. Each query corresponds to a spatial arrangement of several local topographic codes (or tags), which are in themselves too primitive and common to be informative about shape. All the discriminating power derives from relative angles and distances among the tags. The important attributes of the queries are a natural partial ordering corresponding to increasing structure and complexity; semi-invariance, meaning that most shapes of a given class will answer the same way to two queries that are successive in the ordering; and stability, since the queries are not based on distinguished points and substructures.
No classifier based on the full feature set can be evaluated, and it is impossible to determine a priori which arrangements are informative. Our approach is to select informative features and build tree classifiers at the same time by inductive learning. In effect, each tree provides an approximation to the full posterior where the features chosen depend on the branch that is traversed. Due to the number and nature of the queries, standard decision tree construction based on a fixed-length feature vector is not feasible. Instead we entertain only a small random sample of queries at each node, constrain their complexity to increase with tree depth, and grow multiple trees. The terminal nodes are labeled by estimates of the corresponding posterior distribution over shape classes. An image is classified by sending it down every tree and aggregating the resulting distributions.
The method is applied to classifying handwritten digits and synthetic linear and nonlinear deformations of three hundred [Formula: see text] symbols. State-of-the-art error rates are achieved on the National Institute of Standards and Technology database of digits. The principal goal of the experiments on [Formula: see text] symbols is to analyze invariance, generalization error and related issues, and a comparison with artificial neural networks methods is presented in this context.
[Figure: see text]
Where feature points are used in real-time frame-rate applications, a high-speed feature detector is necessary. Feature detectors such as SIFT (DoG), Harris and SUSAN are good methods which yield high quality features, however they are too computationally intensive for use in real-time applications of any complexity. Here we show that machine learning can be used to derive a feature detector which can fully process live PAL video using less than 7% of the available processing time. By comparison neither the Harris detector (120%) nor the detection stage of SIFT (300%) can operate at full frame rate.
Clearly a high-speed detector is of limited use if the features produced are unsuitable for downstream processing. In particular, the same scene viewed from two different positions should yield features which correspond to the same real-world 3D locations [1]. Hence the second contribution of this paper is a comparison corner detectors based on this criterion applied to 3D scenes. This comparison supports a number of claims made elsewhere concerning existing corner detectors. Further, contrary to our initial expectations, we show that despite being principally constructed for speed, our detector significantly outperforms existing feature detectors according to this criterion.
In the context of minimally invasive cardiac surgery, active vision-based motion compensation schemes have been proposed for mitigating problems related to physiological motion. However, robust and accurate visual tracking is a difficult task. The purpose of this paper is to present a hybrid tracker that estimates the heart surface deformation using the outputs of multiple visual tracking techniques. In the proposed method, the failure of an individual technique can be circumvented by the success of others, enabling the robust estimation of the heart surface deformation with increased spatial resolution. In addition, for coping with the absence of visual information due to motion blur or occlusions, a temporal heart motion model is incorporated as an additional support for the visual tracking task. The superior performance of the proposed technique compared to existing techniques individually is demonstrated through experiments conducted on recorded images of an in vivo minimally invasive CABG using the DaVinci robotic platform.
Additional material can be found at
http://www.lirmm.fr/~richa/my_files/video_miccai10.flv
and
http://www.lirmm.fr/~richa/my_files/evaluating_tracking_quality.pdf
.
Accurate estimation and tracking of dynamic tissue deformation is important to motion compensation, intra-operative surgical guidance and navigation in minimally invasive surgery. Current approaches to tissue deformation tracking are generally based on machine vision techniques for natural scenes which are not well suited to MIS because tissue deformation cannot be easily modeled by using ad hoc representations. Such techniques do not deal well with inter-reflection changes and may be susceptible to instrument occlusion. The purpose of this paper is to present an online learning based feature tracking method suitable for in vivo applications. It makes no assumptions about the type of image transformations and visual characteristics, and is updated continuously as the tracking progresses. The performance of the algorithm is compared with existing tracking algorithms and validated on simulated, as well as in vivo cardiovascular and abdominal MIS data. The strength of the algorithm in dealing with drift and occlusion is validated and the practical value of the method is demonstrated by decoupling cardiac and respiratory motion in robotic assisted surgery.
This paper presents an online feature selection mechanism for evaluating multiple features while tracking and adjusting the set of features used to improve tracking performance. Our hypothesis is that the features that best discriminate between object and background are also best for tracking the object. Given a set of seed features, we compute log likelihood ratios of class conditional sample densities from object and background to form a new set of candidate features tailored to the local object/background discrimination task. The two-class variance ratio is used to rank these new features according to how well they separate sample distributions of object and background pixels. This feature evaluation mechanism is embedded in a mean-shift tracking system that adaptively selects the top-ranked discriminative features for tracking. Examples are presented that demonstrate how this method adapts to changing appearances of both tracked object and scene background. We note susceptibility of the variance ratio feature selection method to distraction by spatially correlated background clutter and develop an additional approach that seeks to minimize the likelihood of distraction.
In this paper, we compare the performance of descriptors computed for local interest regions, as, for example, extracted by the Harris-Affine detector. Many different descriptors have been proposed in the literature. It is unclear which descriptors are more appropriate and how their performance depends on the interest region detector. The descriptors should be distinctive and at the same time robust to changes in viewing conditions as well as to errors of the detector. Our evaluation uses as criterion recall with respect to precision and is carried out for different image transformations. We compare shape context, steerable filters, PCA-SIFT, differential invariants, spin images, SIFT, complex filters, moment invariants, and cross-correlation for different types of interest regions. We also propose an extension of the SIFT descriptor and show that it outperforms the original method. Furthermore, we observe that the ranking of the descriptors is mostly independent of the interest region detector and that the SIFT-based descriptors perform best. Moments and steerable filters show the best performance among the low dimensional descriptors.
In robotically assisted laparoscopic surgery, soft-tissue motion tracking and structure recovery are important for intraoperative surgical guidance, motion compensation and delivering active constraints. In this paper, we present a novel method for feature based motion tracking of deformable soft-tissue surfaces in totally endoscopic coronary artery bypass graft (TECAB) surgery. We combine two feature detectors to recover distinct regions on the epicardial surface for which the sparse 3D surface geometry may be computed using a pre-calibrated stereo laparoscope. The movement of the 3D points is then tracked in the stereo images with stereo-temporal constrains by using an iterative registration algorithm. The practical value of the technique is demonstrated on both a deformable phantom model with tomographically derived surface geometry and in vivo robotic assisted minimally invasive surgery (MIS) image sequences.
The hazards of surgical smoke are well documented and electrosurgical units (ESUs) are an integral part of surgical practice. The aim of this study was to gauge the opinions of general surgical consultants, specialist registrars and senior theatre nurses in the Wessex Region towards the hazards of ESU smoke.
A literature search was carried out using Ovid Medline. A questionnaire was sent to 169 consultants, SpRs and nurses in the 14 hospitals across the Wessex Region, exploring current practices, perceived hazards and whether adequate precautions were currently in use.
Only 3 of 98 surgeons used dedicated smoke extractors, despite the fact the majority (72%) felt that, currently, inadequate precautions were taken to protect staff and patients from surgical smoke. There was also uncertainty about the hazards amongst the respondents.
The use of smoke extraction equipment is very limited. Greater awareness of the hazards and available technology to extract fumes from the theatre environment might lead to greater uptake.
Accurate patient registration and referencing is a key element in navigated surgery. Unfortunately all existing methods are either invasive or very time consuming. We propose a fully non-invasive optical approach using a tracked monocular endoscope to reconstruct the surgical scene in 3D using photogrammetric methods. The 3D reconstruction can then be used for matching the pre-operative data to the intra-operative scene. In order to cope with the near real-time requirements for referencing, we use a novel, efficient 3D point management method during 3D model reconstruction.
The presented prototype system provides a reconstruction accuracy of 0.1 mm and a tracking accuracy of 0.5 mm on phantom data. The ability to cope with real data is demonstrated by cadaver experiments.
The use of vision based algorithms in minimally invasive surgery has attracted significant attention in recent years due to its potential in providing in situ 3D tissue deformation recovery for intra-operative surgical guidance and robotic navigation. Thus far, a large number of feature descriptors have been proposed in computer vision but direct application of these techniques to minimally invasive surgery has shown significant problems due to free-form tissue deformation and varying visual appearances of surgical scenes. This paper evaluates the current state-of-the-art feature descriptors in computer vision and outlines their respective performance issues when used for deformation tracking. A novel probabilistic framework for selecting the most discriminative descriptors is presented and a Bayesian fusion method is used to boost the accuracy and temporal persistency of soft-tissue deformation tracking. The performance of the proposed method is evaluated with both simulated data with known ground truth, as well as in vivo video sequences recorded from robotic assisted MIS procedures.
In earlier work, we proposed treating wide baseline matching of feature points as a classification problem, in which each class corresponds to the set of all possible views of such a point. We used a K-mean plus Nearest Neighbor classifier to validate our approach, mostly because it was simple to implement. It has proved effective but still too slow for real-time use. In this paper, we advocate instead the use of randomized trees as the classification technique. It is both fast enough for real-time performance and more robust. It also gives us a principled way not only to match keypoints but to select during a training phase those that are the most recognizable ones. This results in a real-time system able to detect and position in 3D planar, non-planar, and even deformable objects. It is robust to illuminations changes, scale changes and occlusions.
This work presents first in-vivo results of beating heart tracking with a surgical robot arm in off-pump cardiac surgery. The tracking is performed in a 2D visual servoing scheme using a 500 frame per second video camera. Heart motion is measured by means of active optical markers that are put onto the heart surface. Amplitude of the motion is evaluated along the two axis of the image reference frame. This is a complex and fast motion that mainly reflects the influence of both the respiratory motion and the electro-mechanical activity of the myocardium. A model predictive controller is setup to track the two degrees of freedom of the observed motion by computing velocities for two of the robot joints. The servoing scheme takes advantage of the ability of predictive control to anticipate over future references provided they are known or they can be predicted. An adaptive observer is defined along with a simple cardiac model to estimate the two components of the heart motion. The predictions are then fed into the controller references and it is shown that the tracking behaviour is greatly improved.
We present a method for evaluating multiple feature spaces while tracking, and for adjusting the set of features used to improve tracking performance. Our hypothesis is that the features that best discriminate between object and background are also best for tracking the object. We develop an online feature selection mechanism based on the two-class variance ratio measure, applied to log likelihood distributions computed with respect to a given feature from samples of object and background pixels. This feature selection mechanism is embedded in a tracking system that adaptively selects the top-ranked discriminative features for tracking. Examples are presented to illustrate how the method adapts to changing appearances of both tracked object and scene background.
One of the problems tightly linked to endoscopic surgery is the
fact that, because of the narrow field of view, it is sometimes quite
difficult to locate the objects that can be seen through the endoscope.
This is especially true in cardiac surgery, where it is difficult to not
confuse two coronary arteries on a beating heart. We propose a
methodology to achieve coronary localisation by augmented reality on a
robotized stereoscopic endoscope. The method we propose involves five
steps: making a time-variant 3D model of the beating heart using
coronarography and CT-scan or MRI, calibrating the stereoscopic
endoscope, reconstructing the 3D operating field, registering the
operating field surface with the 3D heart model, and adding information
on the endoscopic images by augmented reality. The da Vinci<sup>TM</sup>
surgical system was used for our first experiments with the Cardiac
Surgery team at Hopital Europeen Georges Pompidou, Paris, France
As a first step towards a perceptual user interface, a computer vision color tracking algorithm is developed and applied towards tracking human faces. Computer vision algorithms that are intended to form part of a perceptual user interface must be fast and efficient. They must be able to track in real time yet not absorb a major share of computational resources: other tasks must be able to run while the visual interface is being used. The new algorithm developed here is based on a robust...
In robotic assisted beating heart surgery, motion of the heart surface might be virtually stabilized to let the surgeon work as in on-pump cardiac surgery. Virtual stabilization means to compensate physically the relative motion between instrument tool tip and point of interest on the heart surface, and to offer surgeon a stable visual display of the scene. In this way, motion of the heart must be estimated. This article focus on motion estimation of heart surface. Classical computer vision method has been applied to reconstruct 3D pose of interest point. Experimental results obtained on in vivo images show the estimated motion of heart surface points.
As a first step towards a perceptual user interface, a computer vision color tracking algorithm is developed and applied towards tracking human faces. Computer vision algorithms that are intended to form part of a perceptual user interface must be fast and efficient. They must be able to track in real time yet not absorb a major share of computational resources: other tasks must be able to run while the visual interface is being used. The new algorithm developed here is based on a robust non- parametric technique for climbing density gradients to find the mode (peak) of probability distributions called the mean shift algorithm. In our case, we want to find the mode of a color distribution within a video scene. Therefore, the mean shift algorithm is modified to deal with dynamically changing color probability distributions derived from video frame sequences. The modified algorithm is called the Continuously Adaptive Mean Shift (CAMSHIFT) algorithm. CAMSHIFT's tracking accuracy is compared against a Polhemus tracker. Tolerance to noise, distractors and performance is studied. CAMSHIFT is then used as a computer interface for controlling commercial computer games and for exploring immersive 3D graphic worlds.
The technology for building knowledge-based systems by inductive inference from examples has been demonstrated successfully in several practical applications. This paper summarizes an approach to synthesizing decision trees that has been used in a variety of systems, and it describes one such system, ID3, in detail. Results from recent studies show ways in which the methodology can be modified to deal with information that is noisy and/or incomplete. A reported shortcoming of the basic algorithm is discussed and two means of overcoming it are compared. The paper concludes with illustrations of current research directions.
Local motion on the beating heart is investigated in the context of minimally invasive robotic surgery. The focus lies on the motion remaining in the mechanically stabilised field of surgery of the heart. Motion is detected by tracking natural landmarks on the heart surface in 2D video images. An appropriate motion model is presented with a discussion of its degrees of freedom and a trajectory analysis of its parameters.
Flexible endoscopes are used in many diagnostic and interventional procedures. Physiological motions may render the physicians task very difficult to perform. Assistance could be achieved by using motorized endoscopes and real-time visual tracking algorithm to automatically follow a selected target. In order to control the motors, one needs to have an accurate estimation of the motion of the target in the endoscopic view, which requires an efficient tracking algorithm. In this paper, we compare existing tracking algorithms on various in vivo targets in order to assess their behavior under different conditions. The study shows that several issues have to be overcome by tracking algorithms in in vivo environment like illumination change and forward/backward motions of the target.
With the advent of new applications in cardiac robotic-assisted minimally invasive surgery (MIS), a demand for the design of efficient motion compensation systems was created. In this context, vision-based techniques seem to be a practical way to retrieve the motion of the beating heart since they do not require the introduction of additional sensors in the limited workspace. In this paper, we propose an efficient method for tracking the heart surface which incorporates two novelties. The first is a thin-plate splines (TPS) parametric model for the heart surface deformation that allows us to better track regions of the heart surface with little texture information which undergo large non-rigid deformations. The second novelty is the incorporation of a performing illumination compensation algorithm to cope with arbitrary illumination changes and increase tracking robustness.We also extend this framework for 3D tracking, to enable full compensation of the heart motion. Extensive experiments conducted on in-vivo and ex-vivo heart images attest the notable performance of the algorithm.
Flexible endoscopes are used in many surgical procedures and diagnostic exams, like in gastroscopy or coloscopy. They have also been used recently for new surgical procedures using natural orifices called NOTES. While these procedures are very promising for the patients, they are quite awkward for the surgeons. The flexible endoscope allows the access to operating areas which are not easily reachable, with small or no incisions; but the manipulation of the system is complex. In order to help the practicians during NOTES or classical interventions with flexible endoscopes, we propose to motorize the system so as to partially robotize the movements. This paper presents the problems in the use of the flexible endoscope and explains how the system can be used to stabilize the endoscope on an area of interest despite physiological motions and therefore to improve the manipulation of the system.
In this paper we present a new monochromatic pattern for a robust structured light coding based on the spatial neighborhood scheme using the M-array approach. The proposed pattern is robust as it allows a high error rate characterized by an average Hamming distance higher than 6. We tackle the design problem with the definition of a small set of symbols associated to simple geometrical primitives. One of these primitives embeds the local orientation of the pattern. This is helpful while performing the search for the relevant neighborhood during the decoding process. The aim of this work is to use this pattern for the real-time 3-D reconstruction of dynamic scenes, particularly in endoscopic surgery, with fast and reliable detection and decoding stages. Ongoing results are presented to assess both the capabilities of the proposed pattern and the reliable decoding algorithm with projections onto simple 3-D scenes and onto internal structures of a pig abdomen.
This paper presents a method for reconstructing the 3D shape of an object from its endoscope image based on image shading. The primary problem is that the endoscope has a light source near the object surface. Most of the conventional shape from shading methods assumed that the light source was distant from the object surface and simplified the analysis. To deal with the near light source, we use the configuration of the endoscope that the light source of the endoscope is well approximated by an imaginary point source at the projection center. In addition, we introduce a notion of equal distance contours of the object surface; by propagating the contours using the image shading, we reconstruct the object shape. This is an extension of the Kimmel–Bruckstein algorithm of shape from shading to the endoscope images. Experimental results for real medical endoscope images of the stomach wall show the feasibility of this method and also show its promising availability for morphological analyses of tumors on human inner organs.
Abstract This paper introduces a novel colour-based affine co- variantregiondetector. Ouralgorithmisanextension ofthe maximally stable extremal region (MSER) to colour. The ex- tension to colour is done by looking at successive time-steps of an agglomerative clustering of image pixels. The se- lection of time-steps is stabilised against intensity scalings and image blur by modelling the distribution of edge mag- nitudes. The algorithm contains a novel edge significance measure based on a Poisson image noise model, which we show performs better than the commonly,used Euclidean distance. We compare our algorithm to the original MSER detector and a competing,colour-based blob feature detec- tor, and show through a repeatability test that our detector performs better. We also extend the state of the art in fea- ture repeatability tests, by using scenes consisting of two planes where one is piecewise transparent. This new test is able to evaluate how stable a feature is against changing backgrounds.
This paper presents a method for extracting distinctive invariant features from images that can be used to perform reliable matching between different views of an object or scene. The features are invariant to image scale and rotation, and are shown to provide robust matching across a substantial range of affine distortion, change in 3D viewpoint, addition of noise, and change in illumination. The features are highly distinctive, in the sense that a single feature can be correctly matched with high probability against a large database of features from many images. This paper also describes an approach to using these features for object recognition. The recognition proceeds by matching individual features to a database of features from known objects using a fast nearest-neighbor algorithm, followed by a Hough transform to identify clusters belonging to a single object, and finally performing verification through least-squares solution for consistent pose parameters. This approach to recognition can robustly identify objects among clutter and occlusion while achieving near real-time performance.
This paper presents a new framework for tracking soft tissue deformation in robotic assisted minimally invasive surgery. The method combines optical feature tracking based on stereo-laparoscope images and a constrained geometrical surface model that deforms with feature motion. This has the advantage of relying on reliable salient feature tracking while embedding underlying constraints on the tissue surface for deriving consistent temporal deformation. The proposed framework is resilient to occlusions and specular highlights. The accuracy and robustness of the proposed method are validated using a phantom heart model with known ground truth. To demonstrate the practical value of the method, example in vivo results are also provided.
The inversion of the normal laparoscopic image around the Y-axis has been shown to facilitate the rate of learning of a laparoscopic task in novice subjects. The aim of this study was to investigate the effect of Y-axis image inversion on the performance of experienced laparoscopic surgeons; this had not been previously investigated.
A total of 16 experienced surgeons,who had already carried out more than 50 operative laparoscopic procedures, and 16 novice participants, who had carried out no procedures, were required to make multiple defined incisions under laparoscopic laboratory conditions within ten 1-minute periods. Participants were randomly allocated to perform the task under either normal or Y-axis inverted imaging conditions (eight experienced surgeons and eight novices in each condition).
Y-axis inversion had a significant detrimental effect on the performance of the surgeons, whilst facilitating the performance of novices. The surgeons however, adapted to the inverted condition rapidly, showing a significant improvement in performance over the ten trials.
The Y-axis-inverted image has a detrimental effect on the performance of experienced surgeons, indicating that they have automated to the "fulcrum effect" of the abdominal wall on instrument manipulation. Y-axis-image inversion was found to facilitate significant learning trends, regardless of the participants' level of experience.
Minimally invasive beating-heart surgery offers substantial benefits for the patient, compared to conventional open surgery. Nevertheless, the motion of the heart poses increased requirements to the surgeon. To support the surgeon, algorithms for an advanced robotic surgery system are proposed, which offer motion compensation of the beating heart. This implies the measurement of heart motion, which can be achieved by tracking natural landmarks. In most cases, the investigated affine tracking scheme can be reduced to an efficient block matching algorithm allowing for realtime tracking of multiple landmarks. Fourier analysis of the motion parameters shows two dominant peaks, which correspond to the heart and respiration rates of the patient. The robustness in case of disturbance or occlusion can be improved by specially developed prediction schemes. Local prediction is well suited for the detection of single tracking outliers. A global prediction scheme takes several landmarks into account simultaneously and is able to bridge longer disturbances. As the heart motion is strongly correlated with the patient's electrocardiogram and respiration pressure signal, this information is included in a novel robust multisensor prediction scheme. Prediction results are compared to those of an artificial neural network and of a linear prediction approach, which shows the superior performance of the proposed algorithms.
The introduction of surgical robots in Minimally Invasive Surgery (MIS) has allowed enhanced manual dexterity through the use of microprocessor controlled mechanical wrists. Although fully autonomous robots are attractive, both ethical and legal barriers can prohibit their practical use in surgery. The purpose of this paper is to demonstrate that it is possible to use real-time binocular eye tracking for empowering robots with human vision by using knowledge acquired in situ. By utilizing the close relationship between the horizontal disparity and the depth perception varying with the viewing distance, it is possible to use ocular vergence for recovering 3D motion and deformation of the soft tissue during MIS procedures. Both phantom and in vivo experiments were carried out to assess the potential frequency limit of the system and its intrinsic depth recovery accuracy. The potential applications of the technique include motion stabilization and intra-operative planning in the presence of large tissue deformation.
Performing motion tracking in real-time is an old and recurrent problem in Computer Vision. It has been addressed through a large set of approaches [17], [9], [1], but achieving a high level of robustness is still a challenge, especially with low definition input. In the considered application, tracking the heart motion in endoscopic beating heart sequences, the sensitivity of existing algorithms to visual artifacts and variations in illumination is an issue that calls for improvements. In the prospect of developing a motion compensation architecture for robotically assisted beating heart surgery, we address the problem of visual information retrieval by proposing a new Composite Tracking Algorithm using both template matching and texture analysis. As we will show in this paper, the use of texture characterization of the heart surface improves the overall precision and robustness, in comparison with other prior approaches.
Texture features are widely used for image classification and retrieval. They offer an efficient way to add prior knowledge in image processing, using an appropriate prior parametrization. In this paper, we present the methods we used to select among numerous existing features the most adapted to deal with beating heart tracking, using an experimental database of images and data mining techniques. Then we introduce the way we plan to use this information to reinforce region tracking algorithms in order to track the motion of characteristic landmarks on the heart surface
Stable local feature detection and representation is a fundamental component of many image registration and object recognition algorithms. Mikolajczyk and Schmid (June 2003) recently evaluated a variety of approaches and identified the SIFT [D. G. Lowe, 1999] algorithm as being the most resistant to common image deformations. This paper examines (and improves upon) the local image descriptor used by SIFT. Like SIFT, our descriptors encode the salient aspects of the image gradient in the feature point's neighborhood; however, instead of using SIFT's smoothed weighted histograms, we apply principal components analysis (PCA) to the normalized gradient patch. Our experiments demonstrate that the PCA-based local descriptors are more distinctive, more robust to image deformations, and more compact than the standard SIFT representation. We also present results showing that using these descriptors in an image retrieval application results in increased accuracy and faster matching.
The 3D reconstruction problem from a single endoscope image of a
smooth object is studied in the context of the shape from shading
methods and considering a single light source at the camera projection
center. Based on a curve expansion shape from shading algorithm, a
spherical projection model for the endoscope camera and a dichromatic
model for the surface reflectance, an approach to solve practical
problems, namely the endoscope image distortion and the removal of the
image specular reflection component, is presented. Results obtained from
application of this approach to synthetic and real images are presented
A new approach toward target representation and localization, the central component in visual tracking of nonrigid objects, is proposed. The feature histogram-based target representations are regularized by spatial masking with an isotropic kernel. The masking induces spatially-smooth similarity functions suitable for gradient-based optimization, hence, the target localization problem can be formulated using the basin of attraction of the local maxima. We employ a metric derived from the Bhattacharyya coefficient as similarity measure, and use the mean shift procedure to perform the optimization. In the presented tracking examples, the new method successfully coped with camera motion, partial occlusions, clutter, and target scale variations. Integration with motion filters and data association techniques is also discussed. We describe only a few of the potential applications: exploitation of background information, Kalman tracking using motion models, and face tracking.
No feature-based vision system can work unless good features can be identified and tracked from frame to frame. Although tracking itself is by and large a solved problem, selecting features that can be tracked well and correspond to physical points in the world is still hard. We propose a feature selection criterion that is optimal by construction because it is based on how the tracker works, and a feature monitoring method that can detect occlusions, disocclusions, and features that do not correspond to points in the world. These methods are based on a new tracking algorithm that extends previous Newton-Raphson style search methods to work under affine image transformations. We test performance with several simulations and experiments. 1 Introduction IEEE Conference on Computer Vision and Pattern Recognition (CVPR94) Seattle, June 1994 Is feature tracking a solved problem? The extensive studies of image correlation [4], [3], [15], [18], [7], [17] and sum-of-squared-difference (SSD...
Soft Tissue Tracking for Minimally Invasive Surgery: Learning Local Deformation Online Gaze-Contingent Soft Tissue Deformation Tracking for Minimally Invasive Robotic Surgery, Medical Image Computing and Computer Assisted Intervention
- bullet
- P Mountney
- G.-Z Yang
- G P @bullet Mylonas
- D Stoyanov
- F Deligianni
- A Darzi
- G.-Z Yang
@BULLET Mountney, P., Yang, G.-Z., 2008. Soft Tissue Tracking for Minimally Invasive Surgery: Learning Local Deformation Online, Medical Image Computing and Computer Assisted Intervention, pp. 364-372. @BULLET Mylonas, G.P., Stoyanov, D., Deligianni, F., Darzi, A., Yang, G.-Z., 2005. Gaze-Contingent Soft Tissue Deformation Tracking for Minimally Invasive Robotic Surgery, Medical Image Computing and Computer Assisted Intervention, pp. 843-850. @BULLET Noce, A., Triboulet, J., Poignet, P., 2007. Efficient Tracking of the Heart Using Texture, Annual International Conference of the IEEE Engineering in Medicine and Biology Society, pp. 4480 – 4483.
An Iterative Image Registration Technique with an Application to Stereo Vision A performance evaluation of local descriptors
- bullet
- D G Lowe
- B @bullet Lucas
- T Kanade
- bullet
- N Masson
- F Nageotte
- P Zanne
- M D Mathelin
- bullet
- K Mikolajczyk
- C Schmid
@BULLET Lowe, D.G., 2004. Distinctive Image Features from Scale-Invariant Keypoints. International Journal of Computer Vision 60, 91-110. @BULLET Lucas, B., Kanade, T., 1981. An Iterative Image Registration Technique with an Application to Stereo Vision, International Joint Conference on Artificial Intelligence, pp. 674-679. @BULLET Masson, N., Nageotte, F., Zanne, P., Mathelin, M.d., 2009. In Vivo Comparison of Real-time Tracking Algorithms for Interventional Flexible Endoscopy, International Symposium on Biomedical Imaging, pp. 1350-1353. @BULLET Mikolajczyk, K., Schmid, C., 2005. A performance evaluation of local descriptors. IEEE Transactions on Pattern Analysis and Machine Intelligence 27, 1615-1630. @BULLET Mountney, P., Lo, B.P.L., Thiemjarus, S., Stoyanov, D., Yang, G.-Z., 2007. A Probabilistic Framework for Tracking Deformable Soft Tissue in Minimally Invasive Surgery, Medical Image Computing and Computer Assisted Intervention, pp. 34-41.
Simulated Experiment with Known Ground Truth To quantitatively evaluate the performance of the proposed method with respect to deformation, a simulation was performed to generate synthetic data. A laparoscopic image of 6
- References Albitar
- C Graebling
- P Doignon
Simulated Experiment with Known Ground Truth To quantitatively evaluate the performance of the proposed method with respect to deformation, a simulation was performed to generate synthetic data. A laparoscopic image of 6. References @BULLET Albitar, C., Graebling, P., Doignon, C., 2007. Robust Structured Light Coding for 3D Reconstruction, International Conference on Computer Vision, pp. 1-6. @BULLET Amit, Y., Geman, D., 1997. Shape Quantization and Recognition with Randomized Trees, Neural Computation, pp. 1545–1588.
Motion Estimation in Beating Heart Surgery Physiological motion rejection in flexible endoscopy using visual servoing, International Conference on Robotics and Automation Induction of decision trees
- bullet
- T Okatani
- K Deguchi
- bullet
- T Ortmaier
- M Gröger
- D H Boehm
- V Falk
- G Hirzinger
- bullet
- L Ott
- P Zanne
- F Nageotte
- M Mathelin
- J Gangloff
- C @bullet Quartucci Forster
- H Tozzi
@BULLET Okatani, T., Deguchi, K., 1997. Shape Reconstruction from an Endoscope Image by Shape from Shading Technique for a Point Light Source at the Projection Centre. Computer Vision and Image Understanding 66, 119-131. @BULLET Ortmaier, T., Gröger, M., Boehm, D.H., Falk, V., Hirzinger, G., 2005. Motion Estimation in Beating Heart Surgery. IEEE Transactions on Biomedical Engineering 52, 1729-1740. @BULLET Ott, L., Zanne, P., Nageotte, F., Mathelin, M.d., Gangloff, J., 2008. Physiological motion rejection in flexible endoscopy using visual servoing, International Conference on Robotics and Automation, pp. 2928-2933. @BULLET Quartucci Forster, C., H,. Tozzi, C,L,., 2000. Towards 3D reconstruction of endoscope images using shape from shading, Brazilian Symposium on Computer Graphics and Image Processing, pp. 90-96. @BULLET Quinlan, J.R., 1986. Induction of decision trees. Machine Learning 1. @BULLET Richa, R., Bó, A.P., Poignet, P., 2010. Robust 3D Visual Tracking for Robotic-assisted Cardiac Interventions, Medical Image Computing and Computer Assisted Intervention, pp. 267-274. @BULLET Richa, R., Poignet, P., Liu, C., 2008. Deformable motion tracking of the heart surface, International Conference on Intelligent Robots and Systems, pp. 3997-4003. @BULLET Rosten, E., Drummond, T., 2006. Machine learning for high-speed corner detection, European Conference on Computer Vision, pp. 430–443.