An automated apparatus for presenting depth-rotated three-dimensional objects in human and animal object recognition research

For practical reasons, research on the recognition of objects from different viewpoints has relied almost exclusively on the use of two-dimensional representations of three-dimensional objects. We describe an apparatus that enables the presentation of three-dimensional objects in a discrimination learning paradigm. Three chambers positioned on a movable table allow each of two objects to be presented on either the left or the right side; a viewing window exposes only two of the objects at a time. The objects can be arbitrarily designated as either an S+ or an S-. In addition, they can be placed precisely in any arbitrary start position and rotated in depth in 100 steps of 3.6 degrees each. We have successfully used this apparatus to investigate recognition of depth-rotated objects by both pigeons and humans. By varying the stimuli, number of stimulus chambers, and software programs, the apparatus can be used for other types of tasks and to investigate other types of processes.

Carbon nanotube field emitter

Recently, carbon nanotubes (CNTs), possessing excellent properties as field emitters, are attracting considerable attention as electron emitters of a cold cathode. In this review article, field emission phenomena of carbon nanotubes with various morphologies and surfaces (clean surface or adsorbed molecules on it) revealed by field emission microscopy are first described. Then, the main subject of this article, application of CNTs as electron sources in display devices is reviewed. Other electric devices utilizing CNT-field emitters are also presented.

Generalized interactions using virtual tools within the spring framework: cutting

We present schemes for real-time generalized mesh cutting. Starting with the a basic example, we describe the details of implementing cutting on single and multiple surface objects as well as hybrid and volumetric meshes using virtual tools with single and multiple cutting surfaces. These methods have been implemented in a robust surgical simulation environment allowing us to model procedures ranging from animal dissection to cleft lip correction.

Spring: a general framework for collaborative, real-time surgical simulation

We describe the implementation details of a real-time surgical simulation system with soft-tissue modeling and multi-user, multi-instrument, networked haptics. The simulator is cross-platform and runs on various Unix and Windows platforms. It is written in C++ with OpenGL for graphics; GLUT, GLUI, and MUI for user interface; and supports parallel processing. It allows for the relatively easy introduction of patient-specific anatomy and supports many common file formats. It performs soft-tissue modeling, some limited rigid-body dynamics, and suture modeling. The simulator interfaces to many different interaction devices and provides for multi-user, multi-instrument collaboration over the Internet. Many virtual tools have been created and their interactions with tissue have been implemented. In addition, a number of extra features, such as voice input/output, real-time texture-mapped video input, stereo and head-mounted display support, and replicated display facilities are presented.

Risk reduction in craniofacial surgery using computer-based modeling and intraoperative immersion

We present a two-stage concept for risk reduction in craniofacial surgery, consisting of preoperative risk modeling and intraoperative risk reduction. Preoperatively it is important to find and to visualize risk sources in order to minimize them. Our risk model is composed by superimposition of an isotropic risk potential and an anisotropic tissue field constituent. It is being applied to preoperative planning and simulation of craniofacial surgeries, for example to determine an access path with least overall risk value. In the operation room risks arise mainly from the absence of preoperative planning and simulation data in the operation field. We use a see-through head-mounted display to optimize this situation in order to allow the surgeon to maintain accuracy in the whole process of computer aided surgery. Main steps of the intraoperative immersion are optical tracking of the surgeon wearing the head-mounted display and of the patient, registration of preoperatively calculated planning data with the patient and visualization of the data within the glasses.

PC-compatible computer-generated stimuli for video-task testing

Whereas computer technology allows for the automation of experimentation and data collection, the process of stimulus production has remained a relatively labor-intensive process. A program for automatic computer generation of novel nonverbal stimuli is described in this paper. The program, STIMGEN, allows menu-driven control over the type and appearance of stimuli. Data are presented in which two monkeys matched to sample with high accuracy using stimuli generated with STIMGEN. These data are interpreted to support the usefulness and value of automatic stimulus generation in a variety of applications.

Three-dimensional distribution of bipolar atrial electrogram voltages in patients with congenital heart disease

Voltage differences might be used to distinguish normal atrial tissue from abnormal atrial tissue. This study was aimed at identifying lowest voltage areas in patients with atrial tachycardia after surgical correction of congenital heart disease and to evaluate if identification of these areas in diseased hearts facilitates selection of critical conduction pathways in reentrant circuits as target sites for catheter ablation. Ten patients (four men, age 39 +/- 15 years) with normal sized atria and atrioventricular reciprocating tachycardia (control group) and ten patients (5 men, 32 +/- 7 years) with congenital heart disease and postoperative atrial tachycardia (CL 281 +/- 79 ms) referred for radiofrequency catheter ablation were studied. Mapping and ablation was guided by a three-dimensional electroanatomic mapping system (CARTO) in all patients. In the control group, voltage maps were constructed during sinus rhythm and during tachycardia to evaluate the voltage distribution. The amplitude of bipolar signals was 1.90 +/- 1.45 mV (0.11-8.12 mV, n = 660) during sinus rhythm and 1.45 +/- 1.66 mV (0.12-5.83 mV, n = 440, P < 0.05) during atrioventricular reciprocating tachycardia. In the study group, the amplitude of 1,962 bipolar signals during tachycardia was 1.01 +/- 1.19 mV (0.04-9.40 mV), which differed significantly from the control group during tachycardia (P < 0.0001). No significant difference in the tachycardia cycle length was found (P < 0.05) between the control and study groups. As the lowest voltage measured in normal hearts was 0.1 mV, this value was used as the upper limit of the lowest voltage areas in the patients with congenital heart disease. These areas were identified by detailed voltage mapping and represented by a gray color. Activation and propagation maps were then used to select critical conduction pathways as target sites for ablation. These sites were characterized by fragmented signals in all patients. Ablation resulted in termination of the tachycardia in eight (80%) of ten patients. Complications were not observed. Identification of the lowest voltage areas using a cut-off value of 0.1 mV in congenital heart disease patients with postoperative atrial reentrant tachycardia facilitated the selection of critical conduction pathways as target sites for ablation.

Measurement of ataxic symptoms with a graphic tablet: standard values in controls and validity in Multiple Sclerosis patients

Aim of our study was to find a specific measure for the intensity of upper limb tremor and other ataxic symptoms in Multiple Sclerosis (MS) patients, and to establish standard values and test quality parameters. Three hundred and forty-two consecutive patients with different symptoms in the upper limbs (upper motor neuron symptoms, cerebellar upper limb ataxia, and/or sensory deficits in the upper limbs) and 140 healthy controls took part in the study. All patients and controls had to trace over a 25 cm high figure '8' on a graphic tablet, to tap with the stylus on the tablet and to perform the nine-hole-peg test (9HPT). Patients were additionally examined using clinical standard scales to classify motor dysfunctions of the upper limbs. One hundred and eighty-nine patients and 27 controls were tested twice to investigate the test reliability. Kinematic analysis of the tablet data was performed by kernel estimators, oscillatory activity by spectral analysis. Total power in the 2--10 Hz band was very specific for ataxia versus other motor symptoms. Tapping and 9HPT could well distinguish patients from controls, and patients with predominant motor neuron or cerebellar symptoms from patients with predominant sensory dysfunctions. Mean drawing error did not differ between motor and sensory dysfunctions. The test--retest reliability was similarly high for both spectral analysis and 9HPT.

Bimanual haptic workstation for laparoscopic surgery simulation

Realistic laparoscopic surgical simulators will require real-time graphic imaging and tactile feedback. Our research objective is to develop a cost-effective haptic workstation for the simulation of laparoscopic procedures for training and treatment planning. The physical station consists of a custom-built frame into which laparoscopic trocars and surgical tools may be attached/inserted and which are continuously adjustable to various positions and orientations to simulate multiple laparoscopic surgical approaches. Instruments inserted through the trocars are attached to end effectors of two haptic devices and interfaced to a high speed PC with fast graphics capability. The haptic device transduces 3D motion of the two manually operated surgical instruments into slave maneuvers in virtual space. The slave instrument tips probe the simulated organ. Simulations currently in progress include: 1) Surface-only renderings, deformation, and haptic interactions with elements in the gall gladder surgical field; 2) Voxel-based simulations of the bulk manipulation of tissue; 3) laparoscopic herniorrhaphy. This system provides force feed-forward from the grasped tools to the contact tissue in virtual space, with deformation of the tissue by the virtual probe, and force feedback from the deformed tissue to the operator's hands.

Collaborated surgical works (surgical planning) in virtual space with tactile sensation between Japan and Germany

Surgeons in Japan and Germany applied tele-virtual surgery and a force feedback device during a hepatectomy simulation. Using this system, surgeons in each country were able to perform various surgical maneuvers upon the same patient. They palpated abdominal skin, made electrical scalpel incisions and widened the incision line by using surgical tools in virtual space. While surgeons performed a virtual operation, the force feedback device conveyed tactile sensations. In each location, the force feedback devices and two graphic workstations of equal capability were employed. As each workstation communicated only event signals through an ISDN (64 Kb) line, it was possible to obtain real time tele-virtual surgery without a large capacity communication infrastructure.

Navigation in medical Internet image databases

The world wide web (WWW) changes common ideas of database access. Hypertext Markup Language allows the simultaneous presentation of information from different sources such as static pages, results of queries from) databases or dynamically generated pages. 'Therefore, the metaphor of the WWW itself as a database was proposed by Mendelzon and Nlilo in 1998. Against this background the techniques of navigation within WWW-databases and the semantic types of their queries has e been analysed. Forty eight image repositories of different types and content, but all concerning medical essence, have been found by search-engines. Many different techniques are offered to enable navigation ranging from simple HTML-link-lists to complex applets. The applets in particular promise an improvement for navigation. Within the meta-information for querying, only ACR- and UMLS-encoding were found, but not standardized vocabularies like ICD10 or Terminologia Anatomica. UMLS especially shows that a well defined thesaurus can improve navigation. However, of the analysed databases only the UMLS 'metathesaurus' is currently implemented without providing additional navigation support based on the UMLS 'semantic network'. Including the information about relationships between the concepts of the metathesaurus or using the UMLS semantic network could provide a much easier navigation within a network of concepts pointing to multimedia files stored somewhere in the WWW.

A prototype haptic lumbar puncture simulator

Lumbar punctures (LP) are complex, precise procedures done to obtain cerebro-spinal fluid from a patient for diagnostic purposes. Incorrect techniques resulting from inadequate training or supervision can result in sub-optimal outcomes. As tactile feedback is crucial for a successful lumbar puncture, this procedure serves as an ideal candidate for the development of a haptic training simulator. The intent of this project is to engineer a force feedback LP simulator that provides a safe method of training students (medical students, residents, or trained physicians) for an actual LP procedure on a patient.

Improvement of surgical simulation using dynamic volume rendering

In the last years high efforts have been taken to develop surgical simulators for computer assisted training. However, most of these simulators use simple models of the human's anatomy, which are manually created using modeling software. Nevertheless, medical experts need to perform the training directly with the patient's complex anatomy, which can be received, for example, from digital imaging datasets (CT, MR). A common technique to display these datasets is volume rendering. However, even with high-end hardware only static models can be handled interactively. In surgical simulators a dynamic component is also needed because tissues must be deformed and partially removed. With the combination of springmass models, which are improved by neuro-fuzzy systems, and the recently developed OpenGL Volumizer, surgical simulation using real-time deformable (or dynamic) volume rendering became possible. As an application example the simulator ROBOSIM for minimally invasive neurosurgery is presented.

The first twelve cases of computer assisted periacetabular osteotomy

Image guided freehand navigation of surgical instruments has been applied to the Bernese periacetabular osteotomy, a complex surgical technique for the treatment of dysplastic hips. This navigation system has been introduced into the operating room and has so far been used for 12 patients. Image data from computed tomography (CT) scans are presented in various ways to support the preoperative plan and to provide optimized control of surgical action. Special attention has been paid to the implementation of a sophisticated surgeon-machine interface. This paper describes the features of this novel surgical navigation system and its introduction into the clinical environment.

[3-D localization of cardiac structures in real time]

Radiographic 3-D localisation enables measurements to be made that facilitate the placement of the interventional device during cardiac intervention. To enable the reader to implement the method himself, we describe the computation of 3-D coordinates, acquisition of the imaging and projection data on-line, and the accuracy that can be expected with the method. The 3-D coordinates of a cardiac structure are calculated from the image point coordinates, the projection data and the system constants of a biplane isocentric X-ray unit. Technical imaging errors are corrected a priori. The biplane projection data of a run are acquired on-line and stored in a data base. The image pair of interest is identified automatically from the inscribed run number, and assigned to the projection data from the data base. The target image point is marked on the monitor for 3-D localisation. The accuracy of the method was determined by comparing the calculated and actual cross-sectional points of a centimetre grid imaged in biplane X-ray projections. 3-D localisation took an average of 9.8 +/- 1.2 seconds. Angles and distances were assessed with a standard error of 1.1 degrees and 0.8 mm. The run number is identified correctly in 98.5% of the cases. The mean absolute location error for all points and image pairs was 0.61 +/- 0.32 mm. The accuracy and precision was 0.03 +/- 0.40 mm. Radiographic 3-D localisation can be performed readily and accurately on-line. The results obtained with the method enable interventional decisions to be optimized.

Suitability of the general-purpose graphic printer as an image output device for digital dental x-ray images

OBJECTIVES

The purpose of this study was to evaluate the suitability of the general-purpose graphic printer as an image output device for digital dental x-ray images.

METHODS

The image quality as obtained by a thermal printer and by a dye sublimation printer was investigated. A grid pattern image was used to check parallelism and verticality of lines in each hard copy. A step-wedge image was printed with each printer, and the optical density, gradient, and root mean square granularity were compared. Depiction ability was also compared by using test images, including small signals.

RESULTS

All of the lines were parallel and vertical on hard copies of both printers. The dye sublimation printer showed better results on optical density, gradient, root mean square granularity, and depiction ability.

CONCLUSION

The dye sublimation printer produces images of the depiction ability comparable to the cathode ray tube display and seems suitable as an image output device for digital dental x-ray images.

[VRATS--Virtual Reality Arthroscopy Training Simulator]

The subject of this paper is a highly interactive medical training system for arthroscopic surgery; this is based on computer graphics and virtual reality (VR) techniques and offers an alternative to conventional training methods. To provide the virtual environment, a realistic 3D representation of the knee joint is derived from 2D medical image data. The use of tracking techniques guarantees an intuitive handling of the surgical instruments. The system allows navigation via a virtual camera and interaction with the virtual anatomical structures. First approaches for the simulation of tissue deformation caused by collisions with the instruments are implemented. One important advantage over conventional training systems is the possibility of verifying the training progress. Work is in progress on the realization of tactile feedback with the aim of providing a higher degree of interactive realism.

[Free hand acquisition, reconstruction and visualization of 3D and 4D ultrasound]

3D Ultrasound will find in the next years a wide popularity under the medical imaging applications. The method expands the well-known sonography on the third dimension, therefore it becomes possible to generate spatial 3D views of internal organs. It is further possible to display static (3D) as well as dynamic organs (4D, e.g. pulsating heart). The clarity of the three-dimensional presentation supports very effectively the navigation. In this article we review the upgrading of conventional ultrasound devices on 3D and 4D capabilities, as well as the display of the datasets by corresponding visualisation and filtering approaches.

[Basic principles of data acquisition and data processing for construction of high quality virtual models]

Creating models for virtual reality subdivides into several steps. The aim of the data acquisition is the extraction of nearly isotropic (same solution in all three axes) data sets with low noise content. An approximate isotropy can be achieved by suitable choice of scan parameters. For raw data reconstruction, the application of high-resolution reconstruction algorithms is prohibited due to increased noise. A missing isotropy can computationally be approximated by interpolation. Further noise suppression is achieved by applying filters. Additionally, the contrast of the object for segmentation can be increased by image processing operators. The correct choice of the segmentation method and the editing tools is essential for a precise segmentation with minimal user interaction. Prior to visualization, smoothing the shape of the segmented model (shape-based or morphological interpolation, polygon reduction of wire frame model) further improves the visual appearance of the 3D model.

[3D models for diagnosis and treatment planning in cardiology]

Due to the development new of imaging devices which produce a large number of tomographic slices, advanced techniques for the evaluation of the large amount of data are required. Computer supported extraction of dynamic 3D-models of the patients anatomy from temporal series thus is highly desirable. Since the diagnostician should be able to quickly make sensible decisions based on the models, high accuracy is required within a minimum of time. We present modeling and visualization techniques that are realized within the Cardiac Station. Results for the application of these techniques to cardiac image data demonstrate their usability. Besides giving information about the patients morphology functional parameters can be derived from the data and visualized together with the model. In order to verify the model with the original image data and for the planning of real intervention interaction techniques are presented.

Virtual environments special needs and evaluative methods

This paper presents an overview of the development of the Learning in Virtual Environments programme (LIVE), carried out in special education over the last four years. It is more precisely a project chronology, so that the reader can sense the historical development of the programme rather than giving emphasis to any one particular feature or breakthrough, which are covered in other papers and available through the authors. The project conception in a special school in Nottingham is followed by a description of the development of experiential and communicational virtual learning environments. These are followed, in turn, by the results of our testing programmes which show that experience gained in a virtual environment can transfer to the real world and that their use can encourage self-directed activity in students with severe learning difficulties. Also included is a discussion of the role of virtual learning environments (VLEs) in special education and of its attributes in the context of contemporary educational theory.

Three-dimensional acquisition and visualization of dental arch features from optically digitized models

A method for the acquisition and evaluation of 3D coordinates from anatomically oriented plaster casts is presented, which is based on optical phase shifting profilometry (a fringe projection technique). With the computer-controlled setup, measurements from different views can be combined to obtain a complete three dimensional reconstruction of the model surface. To allow faster evaluation, the result is converted into a range image. From this digital data set the characteristic features like cusp tips can be identified and located semi-automatically. Based on these marks, quantitative values for differences between situation models like local displacements, e.g. during orthodontic treatment, can be determined. The results are visualized as interactively controllable 3D computer graphics, which helps to make spatial relations clearer.