Knowledge about inter-individual shape variations is useful for a variety of applications in orth... more Knowledge about inter-individual shape variations is useful for a variety of applications in orthopaedic surgery. On the one hand, it can be exploited to design surgical devices (guides or implants) that either fit a broader range of patients or are adapted to specific groups of patients (e.g. male or female). On the other hand, in a personalized therapeutic setting it helps reconstructing the patient's anatomy in a robust and automated way from medical image data, or is used for generating patient-specific yet objective surgical reconstruction plans. Statistical shape models (SSM) are capable of capturing the variablity of anatomical shapes contained in a given population. Usually, the statistical analysis is performed on a given set of single-object training shapes (e.g. one bone/organ or a fixed compound of such structures), which are in correspondence with each other. Training instances of joint structures, such as the knee or hip, however, may exhibit different joint postures. Although one may model joint flexibility implicitly by capturing joint motion statistically (Klinder et al, MICCAI 2008; Heap et al., Image Vision Comput. 1996), this approach is beneficial only if relative transformations between individual objects are a statistical property of anatomy, which is, e.g., not the case for knee bending. Therefore, joint posture should be modeled independently by joint-specific degrees of freedoms. We are presenting so-called articulated statistical shape models (ASSM), which model statistical shape variations independently of joint postures. This allows to measure and analyse characteristic shape changes under different joint postures. We see a potential benefit both in biomechanics-based simulation studies for the optimization of surgical devices, as well as a more robust reconstruction of joint structures in medical image data, especially in the presence of low signal-to-noise ratio, pathologies or artifacts.
Processes for the production of aminated sulfurized olefins are improved by performing the reacti... more Processes for the production of aminated sulfurized olefins are improved by performing the reaction in the presence of a tar and charred reaction byproduct reducing amount of an alkali metal or alkaline earth metal compound.
We present a fully automatic method for 3D segmentation of the mandibular bone from CT data. The ... more We present a fully automatic method for 3D segmentation of the mandibular bone from CT data. The method includes an adaptation of statistical shape models of the mandible, the skull base and the midfacial bones, followed by a simultaneous graph-based optimization of adjacent deformable models. The adaptation of the models to the image data is performed according to a heuristic model of the typical intensity distribution in the vincinity of the bone boundary, with special focus on an accurate discrimination of adjacent bones in joint regions. An evaluation of our method based on 18 CT scans shows that a manual correction of the automatic segmentations is not necessary in approx. 60% of the axial slices that contain the mandible.
Although robotic radiosurgery offers a flexible arrangement of treatment beams, generating treatm... more Although robotic radiosurgery offers a flexible arrangement of treatment beams, generating treatment plans is computationally challenging and a time consuming process for the planner. Furthermore, different clinical goals have to be considered during planning and generally different sets of beams correspond to different clinical goals. Typically, candidate beams sampled from a randomized heuristic form the basis for treatment planning. We propose a new approach to generate candidate beams based on deep learning using radiological features as well as the desired constraints. We demonstrate that candidate beams generated for specific clinical goals can improve treatment plan quality. Furthermore, we compare two approaches to include information about constraints in the prediction. Our results show that CNN generated beams can improve treatment plan quality for different clinical goals, increasing coverage from 91.2 to 96.8% for 3,000 candidate beams on average. When including the clin...
According to the American Academy of Implant Dentistry (AAID) the number of patients with bone-an... more According to the American Academy of Implant Dentistry (AAID) the number of patients with bone-anchored dental implants is growing year by year. The goal is to reduce the costs of implant procedures and material, and at the same time ensuring implantation longevity, as well as patient safety and comfort. During surgery, the task is to place an implant into the bone in a stable manner and so that vital structures, such as dental roots, the inferior alveolar nerve, the sinus or major blood vessels are not damaged. One approach to achieve this goal are personalized surgical drill guides. These mechanical components are manufactured specifically for a given patient and can either be placed on top of the bone, gum or teeth. They must be designed accurately to obtain a stable implant position and to prevent the surgeon from damaging vital structures. A prerequisite for such a design is a pre-operative imaging of the relevant structures. Panoramic X-rays (orthopantomograms) are the predomi...
A high realism of avatars is beneficial for virtual reality experiences such as avatar-mediated c... more A high realism of avatars is beneficial for virtual reality experiences such as avatar-mediated communication and embodiment. Previous work, however, suggested that the usage of realistic virtual faces can lead to unexpected and undesired effects, including phenomena like the uncanny valley. This work investigates the role of photographic and behavioral realism of avatars with animated facial expressions on perceived realism and congruence ratings. More specifically, we examine ratings of photographic and behavioral realism and their mismatch in differently created avatar faces. Furthermore, we utilize these avatars to investigate the effect of behavioral realism on perceived congruence between video-recorded physical person’s expressions and their imitations by the avatar. We compared two types of avatars, both with four identities that were created from the same facial photographs. The first type of avatars contains expressions that were designed by an artistic expert. The second ...
This study’s objective was the generation of a standardized geometry of the healthy nasal cavity.... more This study’s objective was the generation of a standardized geometry of the healthy nasal cavity. An average geometry of the healthy nasal cavity was generated using a statistical shape model based on 25 symptom-free subjects. Airflow within the average geometry and these geometries was calculated using fluid simulations. Integral measures of the nasal resistance, wall shear stresses (WSS) and velocities were calculated as well as cross-sectional areas (CSA). Furthermore, individual WSS and static pressure distributions were mapped onto the average geometry. The average geometry featured an overall more regular shape that resulted in less resistance, reduced WSS and velocities compared to the median of the 25 geometries. Spatial distributions of WSS and pressure of the average geometry agreed well compared to the average distributions of all individual geometries. The minimal CSA of the average geometry was larger than the median of all individual geometries (83.4 vs. 74.7 mm²). The...
Successful functional surgery on the nasal framework requires reliable and comprehensive diagnosi... more Successful functional surgery on the nasal framework requires reliable and comprehensive diagnosis. In this regard, the authors introduce a new methodology: Digital Analysis of Nasal Airflow (diANA). It is based on computational fluid dynamics, a statistical shape model of the healthy nasal cavity and rhinologic expertise. diANA necessitates an anonymized tomographic dataset of the paranasal sinuses including the complete nasal cavity and, when available, clinical information. The principle of diANA is to compare the morphology and the respective airflow of an individual nose with those of a reference. This enables morphometric aberrations and consecutive flow field anomalies to localize and quantify within a patient’s nasal cavity. Finally, an elaborated expert opinion with instructive visualizations is provided. Using diANA might support surgeons in decision-making, avoiding unnecessary surgery, gaining more precision, and target-orientation for indicated operations.
Deep Sea Research Part I: Oceanographic Research Papers
Here, we report on different types of shell pathologies of the enigmatic deep-sea (mesopelagic) c... more Here, we report on different types of shell pathologies of the enigmatic deep-sea (mesopelagic) cephalopod Spirula spirula. For the first time, we apply non-invasive imaging methods to: document trauma-induced changes in shell shapes, reconstruct the different causes and effects of these pathologies, unravel the etiology, and attempt to quantify the efficiency of the buoyancy apparatus. We have analysed 2D and 3D shell parameters from eleven shells collected as beach findings from the Canary Islands (Gran Canaria and Fuerteventura), West-Australia, and the Maldives. All shells were scanned with a nanotomm computer tomograph. Seven shells were likely injured by predator attacks: fishes, cephalopods or crustaceans, one specimen was infested by an endoparasite (potentially Digenea) and one shell shows signs of inflammation and one shell shows large fluctuations of chamber volumes without any signs of pathology. These fluctuations are potential indicators of a stressed environment. Pathological shells represent the most deviant morphologies of a single species and can therefore be regarded as morphological end-members. The changes in the shell volume / chamber volume ratio were assessed in order to evaluate the functional tolerance of the buoyancy apparatus showing that these had little effect.
Deformable model-based approaches to 3D image segmentation have been shown to be highly successfu... more Deformable model-based approaches to 3D image segmentation have been shown to be highly successful. Such methodology requires an appearance model that drives the deformation of a geometric model to the image data. Appearance models are usually either created heuristically or through supervised learning. Heuristic methods have been shown to work effectively in many applications but are hard to transfer from one application (imaging modality/anatomical structure) to another. On the contrary, supervised learning approaches can learn patterns from a collection of annotated training data. In this work, we show that the supervised joint dictionary learning technique is capable of overcoming the traditional drawbacks of the heuristic approaches. Our evaluation based on two different applications (liver/CT and knee/MR) reveals that our approach generates appearance models, which can be used effectively and efficiently in a deformable model-based segmentation framework.
Lecture Notes in Computational Vision and Biomechanics, 2015
Statistical shape models (SSM) describe the shape variability contained in a given population. Th... more Statistical shape models (SSM) describe the shape variability contained in a given population. They are able to describe large populations of complex shapes with few degrees of freedom. This makes them a useful tool for a variety of tasks that arise in computer-aided medicine. In this chapter we are going to explain the basic methodology of SSMs and present a variety of examples, where SSMs have been successfully applied.
We present a novel and computationally efficient method for the detection of meniscal tears in Ma... more We present a novel and computationally efficient method for the detection of meniscal tears in Magnetic Resonance Imaging (MRI) data. Our method is based on a Convolutional Neural Network (CNN) that operates on complete 3D MRI scans. Our approach detects the presence of meniscal tears in three anatomical sub-regions (anterior horn, body, posterior horn) for both the Medial Meniscus (MM) and the Lateral Meniscus (LM) individually. For optimal performance of our method, we investigate how to preprocess the MRI data and how to train the CNN such that only relevant information within a Region of Interest (RoI) of the data volume is taken into account for meniscal tear detection. We propose meniscal tear detection combined with a bounding box regressor in a multi-task deep learning framework to let the CNN implicitly consider the corresponding RoIs of the menisci. We evaluate the accuracy of our CNN-based meniscal tear detection approach on 2,399 Double Echo Steady-State (DESS) MRI scans...
Knowledge about inter-individual shape variations is useful for a variety of applications in orth... more Knowledge about inter-individual shape variations is useful for a variety of applications in orthopaedic surgery. On the one hand, it can be exploited to design surgical devices (guides or implants) that either fit a broader range of patients or are adapted to specific groups of patients (e.g. male or female). On the other hand, in a personalized therapeutic setting it helps reconstructing the patient's anatomy in a robust and automated way from medical image data, or is used for generating patient-specific yet objective surgical reconstruction plans. Statistical shape models (SSM) are capable of capturing the variablity of anatomical shapes contained in a given population. Usually, the statistical analysis is performed on a given set of single-object training shapes (e.g. one bone/organ or a fixed compound of such structures), which are in correspondence with each other. Training instances of joint structures, such as the knee or hip, however, may exhibit different joint postures. Although one may model joint flexibility implicitly by capturing joint motion statistically (Klinder et al, MICCAI 2008; Heap et al., Image Vision Comput. 1996), this approach is beneficial only if relative transformations between individual objects are a statistical property of anatomy, which is, e.g., not the case for knee bending. Therefore, joint posture should be modeled independently by joint-specific degrees of freedoms. We are presenting so-called articulated statistical shape models (ASSM), which model statistical shape variations independently of joint postures. This allows to measure and analyse characteristic shape changes under different joint postures. We see a potential benefit both in biomechanics-based simulation studies for the optimization of surgical devices, as well as a more robust reconstruction of joint structures in medical image data, especially in the presence of low signal-to-noise ratio, pathologies or artifacts.
Processes for the production of aminated sulfurized olefins are improved by performing the reacti... more Processes for the production of aminated sulfurized olefins are improved by performing the reaction in the presence of a tar and charred reaction byproduct reducing amount of an alkali metal or alkaline earth metal compound.
We present a fully automatic method for 3D segmentation of the mandibular bone from CT data. The ... more We present a fully automatic method for 3D segmentation of the mandibular bone from CT data. The method includes an adaptation of statistical shape models of the mandible, the skull base and the midfacial bones, followed by a simultaneous graph-based optimization of adjacent deformable models. The adaptation of the models to the image data is performed according to a heuristic model of the typical intensity distribution in the vincinity of the bone boundary, with special focus on an accurate discrimination of adjacent bones in joint regions. An evaluation of our method based on 18 CT scans shows that a manual correction of the automatic segmentations is not necessary in approx. 60% of the axial slices that contain the mandible.
Although robotic radiosurgery offers a flexible arrangement of treatment beams, generating treatm... more Although robotic radiosurgery offers a flexible arrangement of treatment beams, generating treatment plans is computationally challenging and a time consuming process for the planner. Furthermore, different clinical goals have to be considered during planning and generally different sets of beams correspond to different clinical goals. Typically, candidate beams sampled from a randomized heuristic form the basis for treatment planning. We propose a new approach to generate candidate beams based on deep learning using radiological features as well as the desired constraints. We demonstrate that candidate beams generated for specific clinical goals can improve treatment plan quality. Furthermore, we compare two approaches to include information about constraints in the prediction. Our results show that CNN generated beams can improve treatment plan quality for different clinical goals, increasing coverage from 91.2 to 96.8% for 3,000 candidate beams on average. When including the clin...
According to the American Academy of Implant Dentistry (AAID) the number of patients with bone-an... more According to the American Academy of Implant Dentistry (AAID) the number of patients with bone-anchored dental implants is growing year by year. The goal is to reduce the costs of implant procedures and material, and at the same time ensuring implantation longevity, as well as patient safety and comfort. During surgery, the task is to place an implant into the bone in a stable manner and so that vital structures, such as dental roots, the inferior alveolar nerve, the sinus or major blood vessels are not damaged. One approach to achieve this goal are personalized surgical drill guides. These mechanical components are manufactured specifically for a given patient and can either be placed on top of the bone, gum or teeth. They must be designed accurately to obtain a stable implant position and to prevent the surgeon from damaging vital structures. A prerequisite for such a design is a pre-operative imaging of the relevant structures. Panoramic X-rays (orthopantomograms) are the predomi...
A high realism of avatars is beneficial for virtual reality experiences such as avatar-mediated c... more A high realism of avatars is beneficial for virtual reality experiences such as avatar-mediated communication and embodiment. Previous work, however, suggested that the usage of realistic virtual faces can lead to unexpected and undesired effects, including phenomena like the uncanny valley. This work investigates the role of photographic and behavioral realism of avatars with animated facial expressions on perceived realism and congruence ratings. More specifically, we examine ratings of photographic and behavioral realism and their mismatch in differently created avatar faces. Furthermore, we utilize these avatars to investigate the effect of behavioral realism on perceived congruence between video-recorded physical person’s expressions and their imitations by the avatar. We compared two types of avatars, both with four identities that were created from the same facial photographs. The first type of avatars contains expressions that were designed by an artistic expert. The second ...
This study’s objective was the generation of a standardized geometry of the healthy nasal cavity.... more This study’s objective was the generation of a standardized geometry of the healthy nasal cavity. An average geometry of the healthy nasal cavity was generated using a statistical shape model based on 25 symptom-free subjects. Airflow within the average geometry and these geometries was calculated using fluid simulations. Integral measures of the nasal resistance, wall shear stresses (WSS) and velocities were calculated as well as cross-sectional areas (CSA). Furthermore, individual WSS and static pressure distributions were mapped onto the average geometry. The average geometry featured an overall more regular shape that resulted in less resistance, reduced WSS and velocities compared to the median of the 25 geometries. Spatial distributions of WSS and pressure of the average geometry agreed well compared to the average distributions of all individual geometries. The minimal CSA of the average geometry was larger than the median of all individual geometries (83.4 vs. 74.7 mm²). The...
Successful functional surgery on the nasal framework requires reliable and comprehensive diagnosi... more Successful functional surgery on the nasal framework requires reliable and comprehensive diagnosis. In this regard, the authors introduce a new methodology: Digital Analysis of Nasal Airflow (diANA). It is based on computational fluid dynamics, a statistical shape model of the healthy nasal cavity and rhinologic expertise. diANA necessitates an anonymized tomographic dataset of the paranasal sinuses including the complete nasal cavity and, when available, clinical information. The principle of diANA is to compare the morphology and the respective airflow of an individual nose with those of a reference. This enables morphometric aberrations and consecutive flow field anomalies to localize and quantify within a patient’s nasal cavity. Finally, an elaborated expert opinion with instructive visualizations is provided. Using diANA might support surgeons in decision-making, avoiding unnecessary surgery, gaining more precision, and target-orientation for indicated operations.
Deep Sea Research Part I: Oceanographic Research Papers
Here, we report on different types of shell pathologies of the enigmatic deep-sea (mesopelagic) c... more Here, we report on different types of shell pathologies of the enigmatic deep-sea (mesopelagic) cephalopod Spirula spirula. For the first time, we apply non-invasive imaging methods to: document trauma-induced changes in shell shapes, reconstruct the different causes and effects of these pathologies, unravel the etiology, and attempt to quantify the efficiency of the buoyancy apparatus. We have analysed 2D and 3D shell parameters from eleven shells collected as beach findings from the Canary Islands (Gran Canaria and Fuerteventura), West-Australia, and the Maldives. All shells were scanned with a nanotomm computer tomograph. Seven shells were likely injured by predator attacks: fishes, cephalopods or crustaceans, one specimen was infested by an endoparasite (potentially Digenea) and one shell shows signs of inflammation and one shell shows large fluctuations of chamber volumes without any signs of pathology. These fluctuations are potential indicators of a stressed environment. Pathological shells represent the most deviant morphologies of a single species and can therefore be regarded as morphological end-members. The changes in the shell volume / chamber volume ratio were assessed in order to evaluate the functional tolerance of the buoyancy apparatus showing that these had little effect.
Deformable model-based approaches to 3D image segmentation have been shown to be highly successfu... more Deformable model-based approaches to 3D image segmentation have been shown to be highly successful. Such methodology requires an appearance model that drives the deformation of a geometric model to the image data. Appearance models are usually either created heuristically or through supervised learning. Heuristic methods have been shown to work effectively in many applications but are hard to transfer from one application (imaging modality/anatomical structure) to another. On the contrary, supervised learning approaches can learn patterns from a collection of annotated training data. In this work, we show that the supervised joint dictionary learning technique is capable of overcoming the traditional drawbacks of the heuristic approaches. Our evaluation based on two different applications (liver/CT and knee/MR) reveals that our approach generates appearance models, which can be used effectively and efficiently in a deformable model-based segmentation framework.
Lecture Notes in Computational Vision and Biomechanics, 2015
Statistical shape models (SSM) describe the shape variability contained in a given population. Th... more Statistical shape models (SSM) describe the shape variability contained in a given population. They are able to describe large populations of complex shapes with few degrees of freedom. This makes them a useful tool for a variety of tasks that arise in computer-aided medicine. In this chapter we are going to explain the basic methodology of SSMs and present a variety of examples, where SSMs have been successfully applied.
We present a novel and computationally efficient method for the detection of meniscal tears in Ma... more We present a novel and computationally efficient method for the detection of meniscal tears in Magnetic Resonance Imaging (MRI) data. Our method is based on a Convolutional Neural Network (CNN) that operates on complete 3D MRI scans. Our approach detects the presence of meniscal tears in three anatomical sub-regions (anterior horn, body, posterior horn) for both the Medial Meniscus (MM) and the Lateral Meniscus (LM) individually. For optimal performance of our method, we investigate how to preprocess the MRI data and how to train the CNN such that only relevant information within a Region of Interest (RoI) of the data volume is taken into account for meniscal tear detection. We propose meniscal tear detection combined with a bounding box regressor in a multi-task deep learning framework to let the CNN implicitly consider the corresponding RoIs of the menisci. We evaluate the accuracy of our CNN-based meniscal tear detection approach on 2,399 Double Echo Steady-State (DESS) MRI scans...
Uploads
Papers by Stefan Zachow