WO2009045827A2 - Procédés et systèmes de localisation d'outils et de repérage d'outils d'instruments robotiques dans des systèmes chirurgicaux robotiques - Google Patents

Procédés et systèmes de localisation d'outils et de repérage d'outils d'instruments robotiques dans des systèmes chirurgicaux robotiques Download PDF

Info

Publication number
WO2009045827A2
WO2009045827A2 PCT/US2008/077611 US2008077611W WO2009045827A2 WO 2009045827 A2 WO2009045827 A2 WO 2009045827A2 US 2008077611 W US2008077611 W US 2008077611W WO 2009045827 A2 WO2009045827 A2 WO 2009045827A2
Authority
WO
WIPO (PCT)
Prior art keywords
image
robotic
tool
information
robotic instrument
Prior art date
Application number
PCT/US2008/077611
Other languages
English (en)
Other versions
WO2009045827A3 (fr
Inventor
Wenyi Zhao
Christopher J. Hasser
William C. Nowlin
Brian D. Hoffman
Original Assignee
Intuitive Surgical, Inc.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from US11/865,015 external-priority patent/US8147503B2/en
Priority claimed from US11/865,016 external-priority patent/US8073528B2/en
Priority claimed from US11/865,014 external-priority patent/US8108072B2/en
Application filed by Intuitive Surgical, Inc. filed Critical Intuitive Surgical, Inc.
Publication of WO2009045827A2 publication Critical patent/WO2009045827A2/fr
Publication of WO2009045827A3 publication Critical patent/WO2009045827A3/fr

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
    • A61B34/70—Manipulators specially adapted for use in surgery
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
    • A61B34/30—Surgical robots
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
    • A61B34/30—Surgical robots
    • A61B34/37—Master-slave robots
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/255—Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
    • A61B34/20—Surgical navigation systems; Devices for tracking or guiding surgical instruments, e.g. for frameless stereotaxis
    • A61B2034/2046—Tracking techniques
    • A61B2034/2065—Tracking using image or pattern recognition
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
    • A61B34/30—Surgical robots
    • A61B2034/305—Details of wrist mechanisms at distal ends of robotic arms
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B90/00—Instruments, implements or accessories specially adapted for surgery or diagnosis and not covered by any of the groups A61B1/00 - A61B50/00, e.g. for luxation treatment or for protecting wound edges
    • A61B90/36—Image-producing devices or illumination devices not otherwise provided for
    • A61B90/37—Surgical systems with images on a monitor during operation
    • A61B2090/372—Details of monitor hardware
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
    • A61B34/20—Surgical navigation systems; Devices for tracking or guiding surgical instruments, e.g. for frameless stereotaxis
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B90/00—Instruments, implements or accessories specially adapted for surgery or diagnosis and not covered by any of the groups A61B1/00 - A61B50/00, e.g. for luxation treatment or for protecting wound edges
    • A61B90/36—Image-producing devices or illumination devices not otherwise provided for
    • A61B90/361—Image-producing devices, e.g. surgical cameras
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/03—Recognition of patterns in medical or anatomical images
    • G06V2201/034—Recognition of patterns in medical or anatomical images of medical instruments

Definitions

  • the embodiments of the invention relate generally to robots and robotic tools or instruments. More particularly, the embodiments of the invention relate to the acquisition and tracking of the position and orientation of robotic tools or instruments.
  • MIS Minimally invasive surgical
  • robotic e.g., telerobotic
  • An endoscopic camera is typically used to provide images to a surgeon of the surgical cavity so that the surgeon can manipulate robotic surgical tools therein.
  • the robotic surgical tool is not in the field of view of the camera or it is otherwise hidden by tissue or other surgical tools, a surgeon may be left guessing how to move the robotic surgical tool when it is obscured from his view.
  • tissue or organs of interest in a surgical cavity are often obscured from view.
  • a surgeon may have to initially guess the location of an organ of interest within a surgical cavity and search around therein to place the organ and the robotic surgical tools within a field view of the endoscopic camera.
  • optical devices such as light emitting diodes
  • optical devices can interfere with endoscopic surgical procedures and may not provide sufficiently accurate position and orientation information for a minimally invasive surgical system.
  • a magnetic device may be applied to a robotic surgical tool in an attempt to magnetically sense its location.
  • robotic surgical tools are often formed of metal and a magnetic device may not work well due to the interference generated by the movement of metal-tools and electrical motors in a minimally invasive surgical system. Moreover, these may provide only a single clue of the position of a robotic surgical tool.
  • Figure IA is a block diagram of a robotic medical system including a stereo viewer and an image guided surgery (IGS) system with a tool tracking sub-system.
  • Figure IB is a block diagram of a patient side cart including robotic surgical arms to support and move robotic instruments.
  • Figure 2 is a functional block diagram of the video portion of the IGS system to provide a stereo image in both left and right video channels to provide three-dimensional images in a stereo viewer.
  • Figure 3 is a perspective view of a robotic surgical master control console including a stereo viewer and an IGS system with tool tracking sub- system.
  • Figure 4 is a perspective view of the stereo viewer of the robotic surgical master control console.
  • Figure 5A is a perspective view of a sequence of video frames including video images of a robotic medical tool that may be used to perform tool tracking.
  • Figure 5B illustrates different tool positions of a pair of tools in the field of view of a camera based on kenematic information and video image information.
  • Figure 6A is a functional block diagram of a tool tracking architecture and methodology for a robotic system including one or more robotic instruments.
  • Figure 6B is a flow chart of a tool tracking library and its application.
  • Figure 7 is a block diagram illustrating various techniques that may be combined together to meet the challenges in tool tracking.
  • Figure 8 is a functional flow-chart of a tool tracking system.
  • Figure 9A is a figure to illustrate the process of pure image segmentation to localize a tool within an image.
  • Figure 9B is a figure to illustrate the process of sequence matching and/or model-based synthesis to localize a tool within an image.
  • Figure 9C is a more detailed figure to illustrate the process of model-based synthesis to localize a tool within an image.
  • Figures 10A- 1OB illustrates elements of a state-space model to adaptively fuse robot kinematics information and vision-based information together.
  • Figure 1 IA-C illustrate various image matching techniques that may be used separately or collectively to determine pose information of a tool.
  • Figure 12A is a diagram illustrating adaptive fusion under different viewing conditions.
  • Figure 12B is a diagram illustrating a set up for parallel stereo.
  • Figures 12C-12E are charts illustrating various view geometry statistics that may be used to enhance the performance of the state space model for adaptively fusing information sources together.
  • Figures 13A-13B are diagrams illustrating sequence matching of a feature in a sequence of one or more images from different sources.
  • Figure 14 is a diagram illustrating appearance learning of objects within an image.
  • Figure 15 is a flow chart of the application of tool tracking to image guided surgery.
  • Figure 16 is a perspective view of overlaying a pre-scanned image of tissue onto a depth map of a surgical site to provide image-guided surgery.
  • Figure 17 is a perspective view of a surgical site with an ultrasound tool capturing ultrasound images for overlay onto a display.
  • aspects of the invention include methods, apparatus and integrated systems for tool acquisition (locating) and tool tracking (kinematics-tracking (pose predicting) and full-tracking) of robotic medical tools.
  • the method/system for tool tracking systematically and efficiently integrates robot kinematics and visual information to obtain pose (position/orientation) information, which can be used to obtain a more accurate pose of a robotic surgical tool than robot kinematics or visual information alone, in either a camera coordinate system or a base coordinate system.
  • pose position/orientation
  • Known kinematics transformation can be applied to the pose correction to achieve improved pose in any related coordinate system.
  • a camera coordinate system is a coordinate system based on a chosen camera (for example, (X r , Y r , Z r ) in Figure 12B), or a common reference coordinate system for multiple cameras (for example, (Xs, Ys, Zs) in Figure 12B).
  • tool tracking explores prior available information, such as the CAD models of tools, and dynamically learns the image appearances of the robotic instruments.
  • tool tracking may be markerless so as not to interfere with normal robotic surgical procedures.
  • tool tracking may provide continuous pose information of the robotic instruments including their relationships (e.g. tool A is on top of tool B and hence partially occluding tool B) with other tools so that image-based segmentation of the tools may be avoided.
  • a block diagram of a robotic surgery system 100 is illustrated to perform minimally invasive robotic surgical procedures using one or more robotic arms 158.
  • Aspects of system 100 include telerobotic and autonomously operating features. These robotic arms often support a robotic instrument.
  • a robotic surgical arm e.g., the center robotic surgical arm 158C
  • a stereo or three-dimensional surgical image capture device IOIC such as a stereo endoscope (which may be any of a variety of structures such as a stereo laparoscope, arthroscope, hysteroscope, or the like), or, optionally, some other imaging modality (such as ultrasound, fluoroscopy, magnetic resonance imaging, or the like).
  • Robotic surgery may be used to perform a wide variety of surgical procedures, including but not limited to open surgery, neurosurgical procedures (e.g., stereotaxy), endoscopic procedures (e.g., laparoscopy, arthroscopy, thoracoscopy), and the like.
  • a user or operator O generally a surgeon performs a minimally invasive surgical procedure on patient P by manipulating control input devices 160 at a master control console 150.
  • a computer 151 of the console 150 directs movement of robotically controlled endoscopic surgical instruments 101A-101C via control lines 159, effecting movement of the instruments using a robotic patient-side system 152 (also referred to as a patient-side cart).
  • the robotic patient-side system 152 includes one or more robotic arms 158.
  • the robotic patient-side system 152 includes at least three robotic surgical arms 158A-158C (generally referred to as robotic surgical arms 158) supported by corresponding positioning setup arms 156.
  • the central robotic surgical arm 158C may support an endoscopic camera 101C.
  • the robotic surgical arms 158 A and 158B to the left and right of center may support robotic instruments 101A and 101B, respectively, that may manipulate tissue.
  • Robotic instruments are generally referred to herein by the reference number 101.
  • Robotic instruments 101 may be any instrument or tool that couples to a robotic arm that can be manipulated thereby and can report back kinematics information to the robotic system.
  • Robotic instruments include, but are not limited to, surgical tools, medical tools, bio-medical tools, and diagnostic instruments (ultrasound, computer tomography (CT) scanner, magnetic resonance imager (MRI)).
  • the robotic patient-side system 152 includes a positioning portion and a driven portion.
  • the positioning portion of the robotic patient-side system 152 remains in a fixed configuration during surgery while manipulating tissue.
  • the driven portion of the robotic patient-side system 152 is actively articulated under the direction of the operator O generating control signals at the surgeon's console 150 during surgery.
  • the driven portion of the robotic patient-side system 152 may include, but is not limited or restricted to robotic surgical arms 158A-158C.
  • the instruments 101, the robotic surgical arms 158A-158C, and the set up joints 156,157 may include one or more displacement transducers, positional sensors, and/or orientational sensors 185,186 to assist in acquisition and tracking of robotic instruments. From instrument tip to ground (or world coordinate) of the robotic system, the kinematics information generated by the transducers and the sensors in the robotic patient-side system 152 is reported back to the robotic system and a tool tracking and image guided surgery (IGS) system 351.
  • IGS tool tracking and image guided surgery
  • the positioning portion of the robotic patient-side system 152 that is in a fixed configuration during surgery may include, but is not limited or restricted to set-up arms 156.
  • Each set-up arm 156 may include a plurality of links and a plurality of joints. Each set-up arm may mount via a first set-up-joint 157 to the patient side system 152.
  • An assistant A may assist in pre-positioning of the robotic patient-side system 152 relative to patient P as well as swapping tools or instruments 101 for alternative tool structures, and the like, while viewing the internal surgical site via an external display 154.
  • the external display 154 or another external display 154 may be positioned or located elsewhere so that images of the surgical site may be displayed to students or other interested persons during a surgery. Images with additional information may be overlaid onto the images of the surgical site by the robotic surgical system for display on the external display 154.
  • the robotic patient-side system 152 comprises a cart column 170 supported by a base 172.
  • One or more robotic surgical arms 158 are respectively attached to one or more set-up arms 156 that are a part of the positioning portion of robotic patient-side system 152.
  • the cart column 170 includes a protective cover 180 that protects components of a counterbalance subsystem and a braking subsystem (described below) from contaminants.
  • each robotic surgical arm 158 is used to control robotic instruments 101 A-IOlC. Moreover, each robotic surgical arm 158 is coupled to a set-up arm 156 that is in turn coupled to a carriage housing 190 in one embodiment of the invention, as described below with reference to Figure 3.
  • the one or more robotic surgical arms 158 are each supported by their respective set-up arm 156, as is illustrated in Figure IB.
  • the robotic surgical arms 158A-158D may each include one or more displacement transducers, orientational sensors, and/or positional sensors 185 to generate raw uncon n ected kinematics data, kinematics datum, and/or kinematics information to assist in acquisition and tracking of robotic instruments.
  • the robotic instruments may also include a displacement transducer, a positional sensor, and/or orientation sensor 186 in some embodiments of the invention.
  • one or more robotic instruments may include a marker 189 to assist in acquisition and tracking of robotic instruments.
  • the stereo endoscopic camera IOIC includes an endoscope 202 for insertion into a patient, a camera head 204, a left image forming device (e.g., a charge coupled device (CCD)) 206L, a right image forming device 206R, a left camera control unit (CCU) 208L, and a right camera control unit (CCU) 208R coupled together as shown.
  • the stereo endoscopic camera IOIC generates a left video channel 220L and a right video channel 220R of frames of images of the surgical site coupled to a stereo display device 164 through a video board 218.
  • a lock reference signal is coupled between the left and right camera control units 208L,208R.
  • the right camera control unit generates the lock signal that is coupled to the left camera control unit to synchronize the left view channel to the right video channel.
  • the left camera control unit 208L may also generates the lock reference signal so that the right video channel synchronizes to the left video channel.
  • the stereo display 164 includes a left monitor 230L and a right monitor 230R.
  • the viewfinders or monitors 230L,230R may be provided by a left display device 402L and a right display device 402R, respectively.
  • the stereo images may be provided in color by a pair of color display devices 402L,402R.
  • Additional details of a stereo endoscopic camera and a stereo display may be found in U.S. Patent No. 5,577,991 entitled "Three Dimensional Vision Endoscope with Position Adjustment Means for Imaging Device and Visual Field Mask" filed on 07/07/1995 by Akui et al; U.S. Patent No.
  • Stereo images of a surgical site may be captured by other types of endoscopic devices and cameras with different structures.
  • a single optical channel may be used with a pair of spatially offset sensors to capture stereo images of the surgical site.
  • the master control console 150 of the robotic surgical system 100 may include the computer 151, a binocular or stereo viewer 312, an arm support 314, a pair of control input wrists and control input arms in a workspace 316, foot pedals 318 (including foot pedals 318A-318B), and a viewing sensor 320.
  • the master control console 150 may further include a tool tracking and image guided surgery system 351 coupled to the computer 151 for providing the tool images and tissue images overlaid on the visible surgical site images.
  • the tool tracking and image guided surgery system 351 may be located elsewhere in the robotic surgical system 100, such as the patient side cart 152 or a separate computer system.
  • the stereo viewer 312 has two displays where stereo three-dimensional images of the surgical site may be viewed to perform minimally invasive surgery.
  • the operator O typically sits in a chair, moves his or her head into alignment with the stereo viewer 312 to view the three-dimensional annotated images of the surgical site.
  • the master control console 150 may include the viewing sensor 320 disposed adjacent the binocular display 312. When the system operator aligns his or her eyes with the binocular eye pieces of the display 312 to view a stereoscopic image of the surgical worksite, the operator's head sets off the viewing sensor 320 to enable the control of the robotic instruments 101.
  • the viewing sensor 320 can disable or stop generating new control signals in response to movements of the touch sensitive handles in order to hold the state of the robotic instruments.
  • the processing required for tool tracking and image guided surgery may be entirely performed using computer 151 given a sufficiently capable computing platform.
  • the arm support 314 can be used to rest the elbows or forearms of the operator O (typically a surgeon) while gripping touch sensitive handles of the control input wrists, one in each hand, in the workspace 316 to generate control signals.
  • the touch sensitive handles are positioned in the workspace 316 disposed beyond the arm support 314 and below the viewer 312. This allows the touch sensitive handles to be moved easily in the control space 316 in both position and orientation to generate control signals.
  • the operator O can use his feet to control the foot-pedals 318 to change the configuration of the surgical system and generate additional control signals to control the robotic instruments 101 as well as the endoscopic camera.
  • the computer 151 may include one or more microprocessors 302 to execute instructions and a storage device 304 to store software with executable instructions that may be used to generate control signals to control the robotic surgical system 100.
  • the computer 151 with its microprocessors 302 interprets movements and actuation of the touch sensitive handles (and other inputs from the operator O or other personnel) to generate control signals to control the robotic surgical instruments 101 in the surgical worksite.
  • the computer 151 and the stereo viewer 312 map the surgical worksite into the controller workspace 316 so it feels and appears to the operator that the touch sensitive handles are working over the surgical worksite.
  • the computer 151 may couple to the tool tracking and image guided surgery system 351 to execute software and perform computations for the elements of the image guided surgery unit.
  • the tool tracking system described herein may be considered as operating in an open- loop fashion if the surgeon operating the master console is not considered part of the system. If the robotic instrument is to be automatically controlled with the tool tracking system, such as in visual servoing systems used to control the pose of a robot's end-effector using visual information extracted from images, the tool tracking system may be considered to be operating in a closed visual-feedback loop.
  • the viewer 312 includes stereo images for each eye including a left image 400L and a right image 400R of the surgical site including any robotic instruments 400 respectively in a left viewfinder 40 IL and a right viewfinder 40 IR.
  • the images 400L and 400R in the viewfinders may be provided by a left display device 402L and a right display device 402R, respectively.
  • the display devices 402L,402R may optionally be pairs of cathode ray tube (CRT) monitors, liquid crystal displays (LCDs), or other type of image display devices (e.g., plasma, digital light projection, etc.) .
  • the images are provided in color by a pair of color display devices 402L,402R; such as color CRTs or color LCDs.
  • three dimensional maps (a depth map with respect to a camera coordinate system or equivalently a surface map of an object with respect to its local coordinate system is a plurality of three-dimensional points to illustrate a surface in three dimensions) of the anatomy, derived from alternative imaging modalities (e.g. CT scan, XRAY, or MRI), may also be provided to a surgeon by overlaying them onto the video images of the surgical site.
  • alternative imaging modalities e.g. CT scan, XRAY, or MRI
  • a right image 410R rendered from a three dimensional map such as from a CT scan may be merged onto or overlaid on the right image 400R being displayed by the display device 402R.
  • a rendered left image 410L is merged into or overlaid on the left image 400L of the surgical site provided by the display device 402L.
  • a stereo image may be displayed to map out organ location or tissue location information in the surgical site to the operator O in the control of the robotic instruments in the surgical site, augmenting the operator' s view of the surgical site with information that may not be directly available or visible by an endoscopic camera in the surgical site.
  • a stereo video endoscopic camera IOIC has been shown and described, a mono video endoscopic camera generating a single video channel of frames of images of the surgical site may also be used in a number of embodiments of the invention. Rendered images can also be overlaid onto the frames of images of the single video channel.
  • Tool tracking has a number of applications in robotic surgical systems.
  • One illustrative application of tool tracking is to automatically control the motion of the endoscopic camera so that a surgeon automatically views regions of interest in a surgical site, thus freeing the surgeon from the camera control task.
  • robotic instruments are tracked so that the endoscopic camera is centered in the field of view of the surgical site.
  • Another illustrative application for accurate tool tracking may be used to move a robotic instrument to reach a surgical target (e.g., a tumor) either automatically or by a surgeon.
  • a surgical target e.g., a tumor
  • GUI graphic user interface
  • a tool tracking system and architecture that fully integrates kinematics information and visual information for robust and accurate tool tracking performance. Results of tool localization and tracking are made accurate and reliable by adaptively combining together robust kinematics information and accurate geometric information derived from video.
  • the tool tracking system performs locating (determining absolute locations/poses with stereo video), tracking (integrating visual and kinematics) and predicting (kinematics while the tool or a portion thereof is not visible) functions.
  • Technical capabilities in the tool tracking system include an analysis-by- synthesis for image matching and a sequential Bayesian approach which fuses together visual and kinematic information.
  • An analysis-by-synthesis capability makes it possible to explore the prior information that is of concern, such as information about the tools and not the tissue and surrounding environment.
  • the basic procedure in analysis-by-synthesis is to synthesize an image based on the model (geometry and texture) and the current pose (position/location and orientation) of a tool and then compare it against real images.
  • the error between the real and synthesized images is the driving force for better estimation of tool pose.
  • appearance-based learning may be applied to update the model for a specific tool.
  • matching may be performed using features, such as edges and/or corners, that are more robust to lighting variation.
  • stereo imaging may be used along with the analysis-by- synthesis (ABS) techniques.
  • ABS analysis-by- synthesis
  • a stereo approach may be provided based on the tool (or some of its parts).
  • stereo and ABS techniques may be applied to a sequence of images (e.g., the same location and different orientations or different locations and different orientation). Sequence-based matching makes the procedure less vulnerable to local minimum in the process of optimization/estimation.
  • a sequential Bayesian approach may be applied to fuse visual information and robot kinematics to efficiently track tools.
  • the states provide zero-order kinematics of the tools (e.g. current position and orientation, or pose), and the first-order kinematics of the tools (e.g. translational and angular velocity).
  • higher-order or lower-order state space model may be adopted.
  • a linear Kalman filtering may be applied when the noise can be approximated as Gaussian.
  • an extended Kalman filtering may be used to filter out noise from observations and state dynamics.
  • tool tracking involves determining a pose of a robotic instrument 101 including its position or location (Xt,Yt,Zt) and its orientation in the camera coordinate system as it moves in, around, and out of the surgical site.
  • a full pose description may not only include the location and orientation of a robotic instrument in a three dimensional space but may further include the pose of an end-effector, if any.
  • Positional information or pose as used herein may be used to refer to one or both the location and orientation of a robotic instrument.
  • a sequence of left video frames 500L within a camera coordinate system and a sequence of right video frames 500R within another camera coordinate system may be used for tool tracking in one embodiment of the invention.
  • a single view with a single sequence of video frames may be used for tool tracking in anther embodiment of the invention.
  • a single video frame may be used for tool tracking in yet another embodiment of the invention providing partially corrected pose estimates.
  • a marker 502 on the robotic instrument 101 may be used to assist in the tool tracking if visible or otherwise sensible.
  • the marker 502 is a painted marker minimally altering the robotic instruments.
  • markerless tool tracking is provided with no modification of the robotic instruments.
  • natural image features of a robotic tool may be detected as natural markers and/or image appearance of the tools and the CAD model of tools may be used to provide tool tracking.
  • video information from an endoscopic camera and kinematics of the robotic arm and robotic instrument are used as cues to determine the pose of the robotic instrument in the surgical site.
  • Kinematics information provided by the surgical system 100 may include kinematic position kf , kinematic orientation kf , kinematic linear velocity kf , and kinematic angular velocity kf of one or more robotic instruments 101.
  • the kinematics information may be the result of movement of the robotic surgical arm 158, the robotic instrument 101, or both the robotic surgical arm 158 and robotic instrument 101 at a given time.
  • the kinematics information provided by the surgical system 100 may also include the kinematic position kf , kinematic orientation kf , kinematic linear velocity k t , and kinematic angular velocity k, of the endoscopic camera to provide a frame of reference.
  • FIG. 6A a functional block diagram of a tool tracking architecture and methodology for a surgical system is illustrated in accordance with embodiments of the invention.
  • the main operational stages of tool tracking are illustrated in the middle column of
  • Figure 6A Key technical capabilities associated with tool tracking are illustrated in the left column of Figure 6A but for operational constraints 601. The end results of the tool tracking methodology are illustrated in the right column of Figure 6A.
  • the key technical components of the methodology and architecture may be further categorized into basic building blocks and supporting blocks.
  • the basic building blocks including image matching 609 and a state-space model 613 that are used to provide efficient tool tracking, each of which are responsive to visual information.
  • the supporting blocks include model-based synthesis 611, adaptive fusion 615, and sequence matching 607 to support the implementation of robust and accurate tool tracking.
  • Adaptive fusion 615 fully explores prior information that may be available, including prior kinematics information and prior visual information.
  • Vision information and kinematics information are typically available with known characteristics.
  • Robot kinematics information is usually stable and often accurate but may drift during long periods of time.
  • Vision-based information is very accurate when it can be reliably estimated. Otherwise vision-based information may be very inaccurate.
  • adaptive fusion is used to obtain accurate information fusion from different sources of information as well as similar sources of information.
  • the vision-based information is known to be accurate then the information fusion is heavily biased towards the vision-based information.
  • robot kinematics information is used over the vision-based information to generate a more robust fusion of information. While the quality of robot kinematics is typically uniform, the quality of vision information in terms of image matching and 3D post estimation varies a lot. View geometry statistics may be used to determine the reliability and accuracy of video-based information.
  • Adaptive fusion may also be used to obtain accurate information fusion from similar sources of information.
  • Model-based synthesis is used herein to generally refer to generation or rendering of a template image for use in subsequent matching operations, and includes full synthesis, geometry only synthesis, and implicit synthesis.
  • Full synthesis is a complete synthesis of an image of the robotic instrument.
  • robotic instrument images are generated from a computer aided design (CAD) model based on its geometry and texture.
  • CAD computer aided design
  • Other prior information (the location/orientation of the model), not necessarily accurate, is presented along with the synthesized robotic instrument images for image matching 609.
  • Geometry-only synthesis is the case where the geometry of the robotic instrument is used to synthesize geometry-only images (e.g., edge images). Texture of the model is not used in geometry-only synthesis.
  • Implicit synthesis is the case where images are not actually synthesized. Instead the model (either geometry or texture or both) is implicitly used to perform image matching.
  • the geometric properties e.g., width, length, shape
  • the geometric relationship among them e.g., markers forming a line
  • sequence matching is where objects or features in a sequence of images captured from one camera view are matched against objects or features in a sequence of images captured from a different camera view. In another embodiment of the invention, sequence matching is where objects or features in a sequence of images from a camera view are matched against objects or features in a sequence of synthesized images.
  • the goal of the tool acquisition stage 604 is to obtain the absolute pose information (location and orientation) of the one or more robotic instruments within the field of view of one or more cameras, such as the stereo endoscopic camera 101C.
  • the tool acquisition stage 604 performs a locating function 614 resulting in the location and orientation of the tool.
  • the goal of the tool tracking stage 606 is to dynamically update the absolute pose (location and orientation) of a moving robotic instrument.
  • the tool tracking stage 606 may perform a full-tracking function 616 or a kinematics -tracking (pose prediction) function 618 respectively resulting in either a full-tracking state when both visual information and robot kinematics information are available or a kinematics -tracking state when visual information is not utilized (e.g., tool outside the field of view or occluded).
  • a full-tracking function 616 or a kinematics -tracking (pose prediction) function 618 respectively resulting in either a full-tracking state when both visual information and robot kinematics information are available or a kinematics -tracking state when visual information is not utilized (e.g., tool outside the field of view or occluded).
  • the mode/stage of the tool tracking system changes from tool acquisition to tool tracking after the tool is initially located within the field of view.
  • the tool tracking system may remain in the tool tracking mode/stage.
  • the tool tracking system may change from a tool tracking mode/stage into a tool acquisition mode/stage if the tool is removed from the field of view and then returns into the field of view.
  • the tool tracking system may optionally begin operation with an initialization procedure if there is only a single tool in the field of view. If additional tools are to be tracked, the optional initialization procedure may be skipped as other tools have been located and tracked. If the tools have no markers, the optional initialization procedure may involve tools moving around in order to obtain a robust localization via sequence matching.
  • an optional initialization of the tool tracking system may occur.
  • Mono or stereo video 603 may be used in the tool tracking system and is initialized to begin generation of digital video frames of image data of the surgical site.
  • Kinematics information 605 may also be used in the tool tracking system during initialization to form an initial pose of the robotic instrument.
  • the kinematics information 605 may include positional information, including angular or linear information for example, from sensors located at various places along a robotic arm and the robotic instrument.
  • the kinematics information 605 may be for both the endoscopic camera and robotic instruments such that the relationship between positional information for the robotic instruments and the camera may be determined.
  • Initialization begins with a single tool in the field of view without any occlusions.
  • the system may be initialized for additional robotic instruments in the field of view. If tools have already been located and tracked and a new tool is being added, the new tool can be initialized by placing it into the field of view with the previously located and tracked tools. If no tool has been located and tracked, each tool may be initialized by placing it within the field of view with all other tools outside the field of view.
  • the robotic instrument being initialized may be placed in the center of the surgical site for optimal estimation across the whole space or as close to stereo endoscopic camera IOIC as possible that will allow for accurate stereo computation. With a markerless system, the robotic instrument may be moved and rotated for reliable sequence matching.
  • the tool tracking system enters a tool acquisition stage/mode in the surgical site.
  • Figure 9B graphically illustrates the tool acquisition stage in a surgical site.
  • Stereo video images 500L,500R of the surgical site are captured by the endoscopic camera 101C, including one or more robotic instruments 101 in the surgical site.
  • Stereo video may be used to obtain an absolute initial pose of the one or more robotic instruments 101 in one embodiment of the invention.
  • mono video may be used with kinematics information to estimate absolute initial pose (position and orientation) of the one or more robotic instruments 101.
  • the one or more robotic instruments 101 may include painted markers 502 to assist in tool acquisition and tool tracking in the surgical site.
  • the tool acquisition stage performs a locating function 614 resulting in the initial pose of the one or more robotic instruments 101 in the surgical site.
  • the methodology goes to block 606.
  • the tool tracking system enters a tool tracking mode or stage in the surgical site.
  • the goal of tool tracking is to update the absolute pose information (location and orientation) based on incremental and/or partial information (visual and robot kinematics).
  • the tool tracking system is at a full-tracking state 616 when visual and kinematics information is available. If a robotic instrument is not visible (e.g., tools inside an organ or occluded by other tools) in the surgical site, the tool tracking system is at a kinematics-tracking state 618 for estimating tool pose.
  • the tool tracking system may transition from tool tracking 606 and return to tool acquisition 604 if a tracked tool gets out of field of view and then comes back into the field of view of the camera.
  • a tool tracking application 650 is executed by a system 351 of the robotic surgical system 100.
  • the video board 218 illustrated in Figure 2 may be a part of the IGS system 351 in order to receive the video images from the endoscopic camera over the surgical site.
  • a kinematics application programming interface (API) 660 provides a software interface to receive the raw kinematics data from the surgical system 100.
  • the kinematics API 660 couples the kinematics information to the tool tracking application 650 and a tool tracking library 652.
  • the raw kinematics data 680 is received by an API streamer thread 658 which provides the physical interface to a communication channel (for example, fiber optic cable or Ethernet, and may buffer the raw kinematics data by storing it into a memory, a hard disk, or other data storage device).
  • the tool tracking library 652 may issue data requests to the API streamer thread 658.
  • a video capture thread 656 is coupled to the endoscopic camera to receive the raw endoscopic video feed 670.
  • the raw video 670 may be mono video of a single channel or stereo video of left and right channels.
  • the video capture thread 656 may buffer the raw video data by storing it into a frame buffer memory, a hard disk, or other data storage device.
  • a video application programming interface (API) 659 provides the software interface to receive the raw video data from the surgical system into the tool tracking system.
  • the tool tracking library 652 may issue data requests to the video capture thread 656.
  • API application programming interface
  • the tool tracking library 652 contains the core functionality of combining kinematics (through kinematics API 660) and video (through video API 659) for accurate tracking of tools.
  • the library also provides application program interface so it can be invoked in a certain way by a customer-designed tool tracking application 650
  • the tool tracking library 652 In response to the video data and the kinematics data, the tool tracking library 652 generates corrected kinematics data for the pose of a robotic instrument.
  • the raw kinematics data is corrected for orientation and position of the tools.
  • the corrected kinematics data may be used in a number of applications, such as image guided surgery.
  • the speed of raw kinematics 680 may be 100 to 200 Hertz (Hz) and the speed of raw video 670 may be 30Hz to 60hz and the speed of tool tracking maybe even slower.
  • the speed of the corrected kinematics 690 should be substantially similar to the speed of the raw kinematics 680 for medical applications.
  • the raw kinematics may be passed through.
  • a correction matrix (rotation and translation) may then be used to correct the raw kinematics information from the tool tracking library.
  • the corrected kinematics 690 may be directly output from the tool tracking library 652 where a correction matrix is applied to the raw kinematics. Either way is feasible because the correction matrix corrects the bias in the raw kinematics and the bias changes slowly, for example, slower than 1 Hz.
  • Figure 6 A illustrates a functional block diagram including operational stages of a tool tracking system.
  • Figure 7 is a block diagram illustrating the challenges of performing tool tracking.
  • Figure 8 graphically illustrates a functional block diagram of a tool tracking system 800.
  • the tool tracking system 800 adaptively fuses visual information and robot kinematics in order to achieve robust, accurate and efficient tool tracking.
  • the unknown full pose of a robotic instrument, at a time instant t is represented as a state s t 805B in a Bayesian state-space model 802.
  • the state-space model 802 may use a plurality of posed states 805A-805C to perform tool tracking in the surgical site.
  • the state-space model 802 may generate the corrected kinematics information 690 of the robotic instrument.
  • a CAD tool model 804 (geometry only or both geometry and texture) is used for synthesizing (explicitly or implicitly) an image under a given pose (i.e. state).
  • the relative robot kinematics k t 605 (where the dot above the k being used to represent that the relative or first-derivative measurements of the kinematics information) between time instances t to t+1 can be coupled into the state-space model 802.
  • Visual information 603 from captured images may be amplified and analyzed by an amplifier/filter 808 to control the influence of visual feedback 809 on the fusion of visual information and kinematics information.
  • the amplifier/filter 808 generally implements how view geometry statistics are applied for adaptive fusion. If stereo images 803 are available, the spatial constraints 807 between left and right image pairs may also be explicitly or implicitly explored to assist in tool tracking.
  • the natural challenges are those imposed by realistic operational scenarios.
  • the technical challenges are those caused by proposed tool tracking algorithms when facing natural challenges.
  • the natural challenges for example include cluttering and occlusion 702, illumination variation and image appearance change 703, and viewing singularity 704.
  • the technical challenges include image segmentation 710 and matching ambiguity 712, for example.
  • the natural challenge of illumination and image appearance 703 is where the same scene changes dramatically along with the motion of directional light sources.
  • the image intensity of the same object can be different, depending on the distance of the object from the lighting source and the angle between the lighting source and local surface normal. This makes image-based processing less reliable.
  • specularities from organs, blood that under directional endo-illumination make image processing more challenging.
  • the natural challenge of viewing singularity 704 may occur when three dimensional geometry information is derived from two dimensional images.
  • Three dimensional geometry information derived from two dimensional images is not reliable when the two dimensional projection of a three dimensional object is degenerated. For example, a three dimensional cylindrical tool is projected onto an image plane as a circle.
  • the natural challenge of scene cluttering and occlusion 702 is the case where there could be more than one robotic instrument in the field of view. Additionally, the robotic instruments may be partially or fully submerged with complex and dynamic background of organ tissues, blood and smoke caused by electro-dissection.
  • the technical challenges include image segmentation 710 and matching ambiguity 712. Moreover while efficiency is of concern, a big technical challenge for tool tracking may be reliability and accuracy under realistic situations.
  • pure image segmentation 710 i.e., segmentation of tools from a 2D image only
  • model based synthesis techniques 722 may be used.
  • model based synthesis 722 a CAD model of a robotic instrument may be used to render a clean tool image as a pattern to match against or search within a limited region constrained by the pose information of tool.
  • Another technical challenge in tool tracking, especially markerless tool tracking, is the matching ambiguity of a pair of images 712, either between left and right images or between real and synthesized images. Fundamentally, many areas in an image look alike and non- corresponding areas of two images may appear to be more alike than two corresponding areas (for example, due to illumination variations), making region-based matching ambiguous. To reduce such ambiguity, sequence matching 728 may be applied where a sequence of images will be matched against another sequence of images. Such a method is useful when we use robust and accurate relative kinematics information k t .
  • the respective sequence of three images 811A-811C [It—i, It, It + i] and the respective sequence of three synthesized images [I s t -i, I S t , I s t+1 ] may be used to determine the unknown states. That is, if we know any one of the three states in the three-state sequence [s t _i, s t , s t+1] , we can obtain other missing states through perfect or errorless relative kinematics.
  • sequence matching 728 can provide a more robust and more reliable matching as the number of unknowns is reduced and the same number of observations (real images) are kept.
  • appearance learning techniques 724 may be used to handle image or appearance changes 703 such as from illumination variations and natural wear of a tool, for example.
  • appearance learning techniques handle appearance changes by training the tool tracking system on image samples of the same tool under different viewing conditions.
  • Appearance learning techniques have been used extensively in object tracking to handle appearance change due to illumination variations. For example, parametric models have been built to handle illumination variations. Appearance learning techniques are further illustrated herein with reference to Figure 14 with the use of face images instead of tool images.
  • adaptive fusion techniques 726 may be used to handle the challenges of singularity of viewing geometry or viewing singularity 704.
  • the technique of adaptive fusion is used to explore the available pose information, i.e., predicted state (before correction) when feeding geometric information derived from video into the Bayesian state-space model 802. More specifically, video-derived information has much less weight when fused with robot kinematics information under such conditions. In a Bayesian state-space model 802, this manifests itself as large noise variance in the observation equation.
  • Adaptive fusion may be used to handle the challenges of singularity of viewing geometry in order to provide robust and accurate kinematics information of the tools in a surgical, medical, or other type of robotic system.
  • Pure image segmentation may be used by a tool tracking algorithm to localize tools. Pure image segmentation of tools from a two dimensional image is straightforward if the tools have distinctive features, such as color marks that may be used to identify a tool. However, operational conditions may make pure image segmentation techniques difficult if not impossible to perform.
  • FIG. 9 A an image has a tool 901 hidden by an occlusion 902.
  • the occlusion 902 is so severe that it breaks key steps (e.g., color- and/or shape-based analysis) of pure image segmentation of the image such that the tool 901 cannot be found therein.
  • the tool shape 90 IS is substantially covered over by the occlusion shape 902S in the image illustrated in Figure 9A.
  • an occlusion can only make it more difficult for pure image segmentation techniques to localize a tool.
  • FIG. 9B the image of the tool 901 is again hidden by an occlusion 902.
  • Techniques of sequence matching and/or model-based-synthesis matching may be used to localize the robotic instruments instead of pure image segmentation. Sequence matching was briefly discussed previously.
  • Model based synthesis uses a priori knowledge regarding kinematics and appearance that may be available for the tools that are being tracked.
  • a CAD model 904A of the tool 901 is used to synthesize an image of the tool given the known or hypothesized pose information for the tool.
  • the pose information for the tool may be determined from kinematics information or otherwise and a posed synthesized image 904B may then be generated.
  • the posed synthesized image 904B of the tool 901 may then be used to perform image matching or an image search within the overall image of a surgical site to find the location of the tool 901 therein even though it may be partially occluded.
  • This technique of tool localization may generally be referred to herein as an analysis-by- synthesis approach.
  • Using the synthesized 904B image as a pattern to search for the tool 901 within an image of the surgical site helps overcome the difficulty of an occlusion 902 that may cover the tool 901. Tool image fragments 901' left over after the occlusion 902 is subtracted from the tool image is sufficient to use to determine tool localization. However if the occlusion 902 completely covers over the tool 901, image analysis alone cannot localize tools.
  • image segmentation may be guided by exploring the available prior kinematics and image information. That is, image segmentation may be constrained to be performed within a limited region of the image of the surgical site based on rough pose information of the tool in response to the prior robot kinematics and the CAD model 904A of the tool 901. This technique of tool localization may generally be referred to herein as aided image segmentation in contrast to pure image segmentation that has no constraints.
  • image synthesis also referred to herein as model- based synthesis
  • image analysis/search are key steps in using analysis-by- synthesis methods for tool localization and tool tracking.
  • the image synthesis 911 and image analysis/search 915 processes may be repeatedly performed in an iterative optimization approach to find the best tool pose parameter in response to a given cost function CF 913.
  • an iterative optimization approach an initial pose hypothesis may be formulated to generate the initial synthesized model tool for computation of an initial cost function.
  • the cost function CF 913 is a function of what corresponding features 912,914 are used for matching and how an image is synthesized during image synthesis 911.
  • a synthesize edge image 904S of the tool may be synthesized during image synthesis 911 in response to the CAD geometry of the CAD model 904A of the tool.
  • a synthesized regular image 904B of the tool may be synthesized during image synthesis 911 in response to the CAD geometry and CAD texture of the CAD model 904A of the tool.
  • the synthesize edge image 904S of the tool may be used to perform image matching with edges in a video image 910A of the tool.
  • the synthesized regular image 904B of the tool may be used to perform image matching with a regular video image 910B of the tool.
  • appearance learning may be used to augment the analysis-by- synthesis process for tool localization and tool tracking.
  • an edge image such as illustrate in video image 910A, is typically robust against lighting variations.
  • I s [x] L(M t [ ⁇ ; (P,a,M g )]) (Eq. 1)
  • x [x, y] is the image coordinate
  • ⁇ is the homogeneous camera geometric projection from three dimensions (3D) into two dimensions (2D).
  • the model texture M t can be mapped to the coordinate of image F[x] as a function of the homogeneous camera geometric projection ⁇ and a combination of tool pose (position P and orientation ⁇ of the tool), and the geometry M of the tool model.
  • the model will be decomposed into triangles, the 3D vertex coordinates of which will be described in a coordinate system attached to the model.
  • the model coordinates will first be transformed to a world coordinate system, before being projected to a 2D display coordinate system by applying the camera model. Once in the 2D display coordinate system, each triangle will be rasterized.
  • the synthesis of the final per- pixel color values may be computed via interpolation of color specified on a per- vertex basis, texture mapping and filtering, and the application of lighting models. (Reference: Computer Graphics: Principals and Practice, by James D. Foley, Andrjes van Dam, et.
  • the function L is a mapping function that maps the model texture M t into the real image/appearance I s [x] because the real image varies with lighting conditions and other factors, such as occlusions.
  • Camera coordinates of a 3D point 93 IP on the tool that maps to x may be represented by [X, Y, Zf .
  • a local coordinate of the 3D point 93 IP on the tool that is internal to the tool model may be represented as [X M ,Y M ,Z M ] T .
  • a transformation T [p ⁇ ) of the local coordinate of the 3D point 93 IP on the tool to the camera coordinate of the tool as a function of the tool pose may be written as
  • T T lP ⁇ [X M ,Y M ,Z M ,l] T (Eq. 2)
  • T [p ⁇ ) is a four by four 3D-to-3D rigid transformation matrix that can be further decomposed into translational and rotational parts.
  • One of the simplest cost functions is a sum of squared differences (SSD) that may be used to compare the synthesized tool image I s with the video images of the surgical site.
  • SSD sum of squared differences
  • an SSD is a simple cost function, it is nonlinear (e.g., higher than quadratic) in terms of the pose parameters due to the camera geometric projection ⁇ that is nonlinear and the mapping function L to map model texture to real image/appearance that varies with lighting conditions that may be non-linear.
  • Minimizing a nonlinear cost function C is a complex optimization problem.
  • the complex optimization problem may be broken up into two different steps to more efficiently minimize the cost function C.
  • the first step entails performing an image matching where the raw image pixels or extracted image features of I are used for matching against those of the respective synthesized tool image I s .
  • the second step involves performing a geometry-only optimization in response to the result of the image matching between the raw image pixels or extracted image features and corresponding ones from the respective synthesized tool image I s .
  • Eq.4 represent the step of finding the corresponding 2D feature points x m from I and 3D points X m on the tool via image matching of I s and I.
  • Eq. 5 represents the geometry-only optimization where optimal 3D-2D mapping T [p ⁇ ) can be found given the matched 2D-3D pairs.
  • the function f() is a nonlinear function in the following form ⁇ ( ⁇ T [p ⁇ ) ) .
  • the image matching and analysis-by-synthesis processes may be incorporated into a sequential framework for fusion of vision and kinematics information to obtain more accurate positional information of a tool than would otherwise be available from each alone.
  • appearance learning techniques may be used to handle image and appearance changes, such as from illumination variations and natural wear of a tool, for example.
  • the appearance variations due to changes in illumination may exhibit illumination subspace/cone phenomena or spherical harmonics for example.
  • Appearance learning techniques generally train the tool tracking system on image samples of the same tool under different viewing conditions. Pose specific learning techniques may be used as well as clustering or manifold learning techniques may be used to train the tool tracking system over a large number of samples.
  • basis images for illumination variations 1401A, 1401B, 1401C may be used to train the tool tracking system to generate one or more synthesized images 1402A- 1402B which are more closely matched to the respective real images 1045A-1405B that may be captured under different lighting conditions.
  • Appearance learning techniques have been used extensively in object tracking to handle appearance change due to illumination variations (reference: G. Hager and P. Belhumeur, "Efficient Region Tracking with Parametric Models of Geometry and Illumination,” IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 20, pp. 1025-1039, 1998).
  • parametric models have been built to handle illumination variations (reference: H. Murase, S.. Nayar, "Learning and Recognition of 3-D Objects from Brightness Images," Proc. AAAI Fall Symposium, Machine Learning in Computer Vision, pp. 25-29, 1993.
  • the next step to obtain more accurate positional information is the fusion of image positional information and kinematics positional information of the robotic instruments.
  • the purpose of information fusion is to provide more robust and/or more accurate positional information for the robotic instruments in the surgical site such that tool tracking information may be applied in various ways to obtain accurate results, e.g., measurements of certain physical entities within a surgical site.
  • Key to successfully fusing information together from similar sources or from different sources is determining how to adjust the contributions of each to the fusion.
  • the contribution of sources to the information fusion may be adjusted in different ways, such as by a winner- take- all or a weighted averaging method, for example.
  • all sources of information should be fused together so that the information fusion constantly provides the best accuracy and the most robust tool tracking.
  • accuracy and robustness in information fusion.
  • the typical practical approaches to information fusion tend to have compromised results.
  • FIG. 10A- 1OB a state-space model is now described to adaptively fuse together robot kinematics information and vision-based information.
  • Both raw robotic kinematics information 1010 of the robotic instrument and vision-based information 1011 can be used to generate the state variables 1000A-1000D.
  • a tool model 804 may be used to synthesize 611 the synthesized images 810 in response to the state variables 1000A-1000D.
  • An image analysis 806, 609 is performed comparing the synthesized images 810 of the robotic instrument with the observed images 1011.
  • Some real-world data analysis tasks involve estimating unknown quantities from given observations. Moreover, a priori knowledge about phenomenon of a number of applications may be available to allow us to formulate Bayesian models involving probability and statistics.
  • the unknown quantities of information fusion at corresponding time instances can be defined as state variables 1000 A-1000D (of the system) and a Markov chain model can be assumed among states at different time instances, then a state-space model may be formed.
  • the state- space model may include 1) a dynamic/state model that relates state variables 1000 A- 100OD at different time instances (t-1, t, t+1, t+2) and 2) an observation model that relates state variables S 1000A-1000D to observations O 1002A-1002D.
  • the state-space model (a discrete version is shown -a continuous version is similar and involves temporal integration) may be described by the following set of mathematical equations,
  • Observation model o t Hs t + w t
  • D an d H are the dynamic matrix and observation matrix respectively.
  • ' and ' are respectively the dynamic noise and the observation noise that have Gaussian distributions an d respectively.
  • C d and C 0 are covariance matrices for dynamic model and observation model respectively.
  • analytic expression that is well-known and widely used is a Kalman filter (see R.E. Kalman, "A New Approach to Linear Filtering and Prediction Problems," Trans of the ASM — Journal of Basic engineering, Vol. 82, pp. 35-45, I960.).
  • EKF Extended Kalman Filtering
  • SMC Sequential Monte Carlo
  • states are represented by posterior probability density function (pdf) and sampling techniques are used to generate the posterior probability density function (pdf).
  • s represents a state vector of the states 1000A-1000D.
  • s may be a vector version of the matrix T ⁇ p ⁇ ] (Eq.1), or a collection of point positions [X,Y,Z] T , etc., that are equivalent in theory but may be chosen based on a particular application.
  • Respective velocity information may be readily added to the state space model by taking the first derivative of the positional information.
  • Position and velocity may be represented mathematically as
  • the orientation and angular velocity of the robotic instrument may be represented mathematically using unit quaternion for its minimal- style representation and operation efficiency as
  • the state vector may be represented mathematically in a covariance matrix as follows: ⁇ ,
  • the filter state-space model may be implemented with extended Kalman filtering
  • Extended Kalman filtering or particle filtering may be used for tool tracking because 1) the dynamic state space model is non-linear due to the quaternion representation, and 2) the observation model is non-linear in the case of using 2D images as observations and linear in the case of using stereo-derived 3D points as observations. In one embodiment of the invention, we adopt the following nonlinear equation to model the transfer of the system from state
  • observation matrix C 0 consists of two parts: one for kinematics (the leading diagonal sub-matrix of Eq. 12) and one sub-matrix, a covariance matrix C 0 v , for vision as follows:
  • the observation equation is linear for the vision part and we can construct the observation covariance matrix C 0 v .
  • the observation covariance matrix for the case of parallel camera setup (Fig. 12B):
  • first-order kinematic information in a number of embodiments of the invention, first-order kinematic information
  • first-order vision observations may be derived by tracking image points (i.e., temporal image matching) to provide an extra constraint.
  • estimated states e.g., tool poses
  • observation vectors may be formed for the vision- based information and the observation equation is non-linear due to the perspective projection of 3D points to 2D points.
  • Kalman filtering provides a perfect sequential solution to a batch least square optimization problem, because of the Markovian assumption that given the present state future states are conditionally independent of the past states.
  • tool tracking in reality is a nonlinear problem and it is often not easy to achieve an accurate solution with a sequential approach such as extended Kalman filtering.
  • extended Kalman filtering To achieve accurate results with extended Kalman filtering given a non-linear system, an iterative bundle adjustment process that combines all observations across multiple states may be used starting from initial results provided by the extended Kalman filtering 802. As a general optimization method, bundle adjustment has wide applications.
  • bundle adjustment can be used for sequential matching to be discussed later.
  • bundle adjustment an optimization technique in photogrammetry, refers to the
  • a batch bundle adjustment may be applied at each time instance or at selected time instances based on all the measurements and state estimates that are available from extended Kalman filtering. Applying a batch bundle adjustment in the beginning of time where the state-space model is applied may be preferred because 1) quick convergence to the correct solution to a non-linear optimization problem is desirous from the beginning, and 2) the computation is efficient because there are only a small number of states and observations available.
  • Image matching is used for incorporating vision information into the tool tracking system and its state space model.
  • Image matching (Eq. 4) in the image analysis steps (Eqs. 4 and 5) finds the corresponding 2D feature image points and 3D points on the tool.
  • Image matching also finds the corresponding image features between the left and right images from a stereo camera.
  • Image matching may be facilitated by sequence matching of a temporal sequence of frames of video images.
  • image matching may be facilitated by using a 2D or 3D model of a tool and a video image.
  • Implementation of the two dimensional image matching may alternately be performed by simple intensity-based SSD (sum of squared difference), feature-based matching (for example, a point, a scale-invariant-feature-transform feature: D. Lowe, "Object recognition from local scale-invariant features," Proc. Int. Conf. Computer Vision, 1999.), or probabilistic matching.
  • a 2D image with a 3D model synthesized image more robust features such as edges can be used.
  • the cost function would be the sum of distance measures from the image edge points to the closest synthesized curves/lines (reference: D. Lowe, "Fitting parameterized three-dimensional models to images," IEEE Trans, on Pattern Analysis and Machine Intelligence, Vol. 13, pp. 441-450.).
  • Image matching can be applied in different scenarios, such as temporal matching of images within sequences of images (Figure HA) or spatial matching of images across two stereo views (Figure HC).
  • real images are used to match against synthesized images in the analysis-by-synthesis approach ( Figure HB).
  • two or more of these image matching techniques may be used together to perform image matching.
  • image matching could be applied to corresponding artificial markers attached to instruments, natural image features or image appearances (e.g., instrument tips).
  • the artificial markers are passive visual markers.
  • FIG. 1 IA temporal image matching of a pair of video images l lOlV-l lOlV is illustrated.
  • the video image 1101V of the actual tool 101 (e.g., see Figure IA) is taken at time t resulting in a tool image 101V at time t.
  • the video image 1101V of the same actual tool 101 is taken at a different time, time t+1, by the same camera resulting in a tool image 101V at time t+1.
  • the camera may be fixed with respect to the robotic surgical system while the tool 101 may move relative to the robotic surgical system.
  • Various aspects of the video images of the tool taken at different times may be used to perform image matching. That is one or more of a matching of markers 1110, a matching of natural features 1111, and/or an appearance matching 1112 may be performed. For example, markers 502V on the tool image 101V in the first video image 1101V of the tool 101 may be compared with the markers 502V on the tool image 101V in the second video image 1101V to help determine new pose information of the actual tool 101. Besides marker image matching, other information may be used determine pose information of the actual tool 101. [0156] Referring now to Figure 1 IB, synthesis image matching of a video image 1101V and a synthesized image HOlS is illustrated.
  • the video image 1101V of the actual tool 101 (e.g., see Figure IA) is taken by a camera resulting in a tool image 101V.
  • the synthesized image 110 IS of the tool 101 is generated by a computer having prior knowledge of the actual tool 101 resulting in a synthesized tool image 10 IS.
  • the pose of synthesized tool image 10 IS in the synthesized image 110 IS of the tool 101 attempts to match the pose of the tool represented by the tool image 101V in the video image 1101V at a moment in time.
  • Various aspects of the video image 1101V and the synthesized image 1101S of the tool may be used to perform image matching. That is, one or more of a matching of markers 1110, a matching of natural features 1111, and/or an appearance matching 1112 may be performed. For example, markers 502V in the first video image 1101V of the tool image 101V may be compared with the synthesized markers 502S on the synthesized tool image 10 IS in the synthesized image 110 IS of the same tool 101 to help determine new pose information of the actual tool 101.
  • the left video image 110 IVL of the actual tool 101 (e.g., see Figure IA) is taken with a camera in a first position with respect to the tool, such as a left side, resulting in a left tool image 10 IVL.
  • the right video image 110 IVR of the same actual tool 101 is taken at a different position with respect to the tool, such as the right side, resulting in a right tool image 101VR.
  • the left video image 1 IOIVL and the right video image 1 IOIVR are captured at substantially the same time by their respective cameras.
  • Various aspects of the left and right video images of the tool may be used to perform image matching.
  • markers 502VL on the left tool image 101 VL in the left video image 1101 VL of the tool 101 may be compared with the markers 502VR on the right tool image IOIVR in the right video image 1 IOIVR of the same tool to determine new pose information of the actual tool 101.
  • these image matching techniques may be combined to generate better pose information for the tool. For example, it is natural to combine temporal image matching (Figure HA) with spatial image matching ( Figure HC) or combine temporal image matching (Figure HA) with synthesis image matching ( Figure HD). However, all three techniques may be used all together or flexibly in various combinations to try and obtain the best pose information for the tool.
  • stereo images are used to construct 3D feature points, and these 3D points can be matched (for example, through the popular iterative closest point algorithm, reference: P. Besel and N, McKay, "A method for registration of 3-D shape," IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 14, pp. 239-256, 1992) against corresponding 3D points, for example, markers, on the robotic instrument.
  • matched for example, through the popular iterative closest point algorithm, reference: P. Besel and N, McKay, "A method for registration of 3-D shape," IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 14, pp. 239-256, 1992
  • corresponding 3D points for example, markers
  • image matching has been described with respect to explicit image synthesis, such as a full image synthesis or a geometry-only image synthesis.
  • image matching may be made using implicit image synthesis. That is, images are not actually synthesized. Rather, a computer aided design (CAD) model of the tools and prior pose information of the tools are implicitly used to facilitate image matching.
  • CAD computer aided design
  • artificial markers may be applied to the tools to assist in image matching as is described below.
  • natural markers or features of a tool may be used to assist in image matching. Thus, the following descriptions apply equally well to the case when no artificial markers are present and features of the tool, .i.e., natural markers, are detected directly from the instrument images.
  • Figure 5B illustrates video images 10 IAV and 10 IBV for a respective pair of tools 10 IA and 10 IB in the field of view 510.
  • Figure 5B further illustrates pose information 101AK and 101BK based on kinematics for the respective tools 10 IA and 10 IB in and around the field of view 510.
  • the video images 10 IAV and 101BV and the pose information 101AK and 101BK for the respective tools 101A and 101B may be adaptively fused together to improve the overall pose information for each.
  • a plurality of marker dots 502A and 502B or other types of markers may be affixed to the respective tools 101A and 101B.
  • Video information of the marker dots 502A' and 502B' may be ascertained from the video images 101AV and 101BV of the respective tools 101A and 101B.
  • the image matching problem is simplified from many dot-to-dot matchings to a single pattern matching.
  • a pattern matching or association is straightforward if one tool is in the field of view. If more that one robotic instrument tool is in the field of view, there are two approaches available to solving the image matching problem that can be used alone or combined together.
  • the robotic kinematics information will not change the spatial arrangements (positional and/or orientational) across instruments, especially in the beginning of surgical operations. For example, if two instruments are arranged to the left and to the right in the camera coordinate system, then the robotic kinematics should represent that arrangement. However, this is not a requirement on the absolute values of the robotic kinematics information.
  • the robotic kinematics information may indicate that the one or both of the robotic surgical tools are outside the field of view. Resolving pattern association ambiguity in matching a first pattern in an image to the tool arranged to the left and a second pattern in the image to the tool arranged to the right can be carried out in either 2D image or 3D space.
  • tool motion is used to resolve the pattern association ambiguity.
  • pattern images are tracked individually through temporal image matching.
  • the motion trajectories of the pattern images are then compared against the kinematic information of each of the tools. From the comparison, the correct association of pattern images to tools can then be made.
  • a pattern association issue for a single tool in the field of view may be directionally ambiguous, such as a line pattern with identical markers that has a directional ambiguity of al80-degree flip.
  • Pattern association for a single tool is not an issue if the pattern consisting of artificial markers or natural markers is unique. For example, if the pattern of markers has directional information embedded such that the markers at each of ends of the pattern are distinctive there is no issue. In one embodiment of the invention, the design of the artificial markers on the tool provides directional information of the marker pattern. [0169] If there is a directional ambiguity for a given tool, there are two approaches to solving the image matching problem, similar to how the pattern associations for multiple tools is handled.
  • the first approach to solving the image matching problem is to use very rough robotic kinematics information to resolve any directional ambiguity.
  • the robotic kinematics should not flip although it may be far away, such as outside the field of view after projection onto 2D image.
  • the second approach to solving the image matching problem is to use motion trajectories of the pattern to remove the directional ambiguity.
  • Image information may have quality issues (matching reliability and accuracy) regarding image-derived 3D information. For example, a viewing- singularity happens when a cylindrical instrument with markers on the shaft is projected to an image of small circle (see
  • Viewing geometry is more than just pure geometry between a camera and a 3D object. Other information related to viewing geometry can impact the quality of image-derived information for fusion.
  • the position of light source 1210A with respect to the camera 121 IA- 121 IB and the 3D object may be of interest to achieve accurate and reliable information.
  • the different poses of the 3D object may change how a light source strikes its features and provide different views.
  • a 3D object may behave differently under the same or different viewing conditions.
  • the tip 1200T and the shaft 1200S of a 3D tool may behave differently under the same or different viewing conditions.
  • the tip 1200T and the shaft 1200S of a 3D tool may behave differently under the same or different viewing conditions.
  • the tip 1200T and the shaft 1200S of a 3D tool may behave differently under the same or different viewing conditions.
  • 1200 may generate different image-derived 3D information with different reliabilities.
  • image-derived 3D information with different reliabilities.
  • the shaft image 122 IS may be good to use while in image 122 IB the shaft image 1222S may be poor to use in forming image-derived 3D information.
  • image-derived 3D information also depends upon the particular image features used. For example, edge-based features are less sensitive to illumination change than intensity-based features.
  • View- geometry statistics are used to represent how good the image-derived information for fusion is when compared to the ground-truth. That is, view-geometry statistics may be used to estimate uncertainty.
  • the view-geometry statistics are used in the Bayesian state-space model for adaptively fusing image information and kinematics information. To be specific, the view geometry statistics may be represented in the covariance matrix of the observation equation (Eq. 13).
  • view-geometry statistics Digitization error/image resolution; feature/algorithm related Image matching error; Distance from object to camera; Angle between object surface normal and line of sight; Illumination and specularity. Based on certain noise assumptions (e.g., independent Gaussian noise), view-geometry statistics for these phenomenon may be computed.
  • noise assumptions e.g., independent Gaussian noise
  • view-geometry statistics for these phenomenon may be computed.
  • Jacobian matrix and x being the mean- subtracted variable as follows; (Eq. 18) x! » B r X,
  • a technique is chosen for matching a pair of image features 1301 and 1302 such as from left and right images respectively in a stereo setting or a real image and a synthesized image.
  • the pair of images features may be matched over one or more sequences of images 1301A-1301E and 1302A-1302E, for example.
  • a rigid sequence matching (where the relative kinematics within each sequence are perfectly known or identical across sequences in the ideal case) may be employed where just one common motion parameter is used to estimate for all pairs between two sequences such as illustrated in Figure 13 A.
  • the sequences 1301 and 1302 have identical motion relationship (relative kinematics) among the images 1301A to 1301E, and 1302A tol302E.
  • a flexible sequence matching (where the relative kinematics within each sequence are known with errors) may be employed as illustrated in Figure 13B.
  • the sequences 1303 and 1302 have a identical motion relationship (relative kinematics) among the images 1303 A tol303E, and 1302A to 1302E.
  • T s, t where C is the cost function of matching features F 1 ; and F 2 ; and X represents the location of the features in either two or three dimensions.
  • the geometric transformation T s l may be in 3D and the features may be in 2D after known camera projection. If relative kinematics Jc 1121 within each sequence are perfectly known, then the geometric transformation equation for sequence matching may become (Eq. 20) where T s u represents transformation from T s , typically a chosen T s l , through relative known kinematics.
  • Sequence matching methods may be applied in many different scenarios, e.g., a single view case (a sequence of real images against a sequence of synthesized images) or a stereo case (matching among two sequences of real images and two sequence of synthesized images).
  • a single view case a sequence of real images against a sequence of synthesized images
  • a stereo case matching among two sequences of real images and two sequence of synthesized images.
  • relative kinematics may not be perfect and may change over time.
  • bundle adjustment based flexible sequence matching may be applied to estimate the optimal parameters to more accurately locate and track tools.
  • a tool tracking system for a robotic instrument has a number of applications.
  • IGS image-guided surgery
  • IGES image-guided endoscopic surgery
  • the basic goal of image guided surgery is to enhance a surgeon's experience by providing real time information derived from single or multiple imaging modalities (e.g., visual, x-ray, computerized topography (CT), magnetic resonance imaging (MRI), ultrasound) during surgery or training/simulation.
  • imaging modalities e.g., visual, x-ray, computerized topography (CT), magnetic resonance imaging (MRI), ultrasound
  • CT computerized topography
  • MRI magnetic resonance imaging
  • Two particular benefits of IGS/IGES are 1) improved visualization for easier on-line diagnostics and 2) improved localization for reliable and precise surgery.
  • Tool tracking is one technology used for IGS/IGES since instruments are used by surgeons to navigate, sense and operate (e.g., diagnostic, cut, suture, ablation etc.) in the areas of interest. That is, the tool tracking system described herein can enable image-guided surgery without significant added operational inconvenience and/or added equipment.
  • Tool tracking may be used to provide automated camera control and guidance to maintain a robotic instrument in the field of view. Tool tracking can also be used to assist the surgeon to move the robotic instrument to reach a tumor either automatically or with a surgeon's assistance. Ultra-sound or pre-scanned images can also be used along with real-time tool tracking. Other applications of tool tracking include graphic user interface that facilities the entrance and re-entrance of the robotic instrument during surgery.
  • Tool tracking can be used to take a number of measurements during surgery as well.
  • tool tracking may be used to measure organ sizes by touching the robotic tool tip at different points of an organ.
  • a pair of robotic tools being tracked can concurrent touch points of the organ and a distance along a line between their tips can be accurately measured with the assistance of tool tracking.
  • tool tracking may be used to construct a 3D model of an organ.
  • the tip of a single robotic tool may be used to touch points across the organ's surface to construct a 3D model of the organ.
  • a perspective view of a surgical site 1700 includes a robotic ultrasound tool 1710.
  • the robotic ultrasound (US) tool 1710 has an attached or integrated ultrasound transducer 1710A.
  • the robotic ultrasound (US) tool 1710 may be used to guide and navigate other instruments to perform various medical or surgical procedures.
  • two dimensional ultrasound images 171 IA may be captured in a two dimensional coordinate system 1703.
  • the two dimensional ultrasound images 171 IA may be translated from the two dimensional coordinate system 1703 into a camera coordinate system 1701.
  • the translated ultrasound images may then be overlaid onto video images of the surgical site 1700 displayed by the stereo viewer 312, such as illustrated by the translated ultrasound images 1711B-1711D in the surgical site 1700 illustrated in Figure 17.
  • Tool tracking may be used to flagpole ultrasound by (1) determining the transformation of the ultrasound images 171 IA from the two dimensional coordinate system 1703 to the local ultrasound coordinate system 1702 in response to ultrasound calibration; (2) at the transducer 1710A, determining the transformation from the ultrasound transducer coordinate system 1702 to the camera coordinate system 1701 by using tool tracking; and then; (3) cascading the transformations together to overlay the ultrasound image in the camera coordinate system 1701 onto the surgical site as illustrated by image 171 IB.
  • the quality of image overlaying depends on tracking accuracy of the robotic ultrasound tool within a camera coordinate system (Fig. 5A) and the ultrasound calibration.
  • Ultrasound calibration is described in the reference "A novel closed form solution for ultrasound calibration," Boctor, E.; Viswanathan, A.; Choti, M.; Taylor, R.H.; Fichtinger, G.; Hager, G., Biomedical Imaging: Nano to Macro, 2004. IEEE International Symposium on Volume , Issue , 15-18 April 2004 Page(s): 527 - 530 Vol. 1.
  • Tool tracking may also be used to generate 3D/volumetric ultrasound images by stacking overlaid 2D images 171 IB-D from the 2D ultrasound transducer 1710A generated by rotating the ultrasound tool 1701 as illustrated by the arrow 1720.
  • tool tracking may be used overlay one or more drop virtual point/marks 1650A-1650B on images of the tissue surface 1600 in the surgical site by using one or multiple tools 1610L, 1610R (for example, a surgical tool or an ultrasound tool) to touch point of interest.
  • tools 1610L, 1610R for example, a surgical tool or an ultrasound tool
  • teaching surgeons can use tools to draw virtual marks to illustrate areas of interest to remote student surgeons on an external display.
  • surgeon can use one type of tracked tool (e.g., an ultrasound tool) to draw marks to indicate regions of interest and then use a different type of tracked tool (e.g., a cautery tool) to operate or perform a surgical or other medical procedure in selected regions of interest.
  • a different type of tracked tool e.g., a cautery tool
  • the tool tracking system described herein may also be used for image-guided interventional radiology (IGIR) along with other sensors.
  • active sensors/cameras e.g. electro-magnetic sensor, or active near Infra-Red illumination plus stereo camera
  • IGS with tool tracking may be used for other medical and surgical procedures for different organs and tissue that are robotically controlled or robotically assisted.
  • a liver or other organ/tissue prior to surgery may be scanned with a computer tomography scanner 1503 to obtain a number of images of the liver such that they may be used to reconstruct a complete 3D volume of a scanned object at block 1504 by a computer as desired.
  • a computer tomography scanner 1503 may be used to obtain a number of images of the liver such that they may be used to reconstruct a complete 3D volume of a scanned object at block 1504 by a computer as desired.
  • E CT computer tomography error
  • a computer tomography volume that includes liver and surrounding area is selected to generate a computer tomographic (CT) segment that contains liver only.
  • CT computer tomographic
  • the CT volume segment is taken to register against a surface/depth map from other imaging modalities, e.g., stereo cameras.
  • the real liver during surgery 1501B may have some biological variation forming an error (E B io) from the prior scan taken by the CT scanner.
  • One or more robotic instrument tips 1510 are inserted into the surgical site.
  • One or more endoscopic cameras 1512 take sequences of images 1513-1514 of the surgical site including the liver 1501B and the robotic instrument tips 1501. These sequences of images 1513-1514 are typically stereo image pairs but could be a sequence of single images from a mono-view.
  • the sequence of images 1513 may be used to determine the depth of surface features to make a surface/depth map 1515.
  • a sequence of surface maps 1515 including a robotic instrument may be analyzed similar to a sequence of images 1513 as described herein to determine a location of the robotic instrument.
  • the surface map 1515 may be overlaid with a model of an organ after surface registration.
  • the surface map 1515 may be further annotated with a map, outline, or other indication of the estimated tumor location 1517.
  • an internal liver tumor that is not visually visible maybe easily visible in a CT scan of the liver.
  • An error between the estimated tumor location and the actual tumor location during surgery may be a sum of errors of the scanning (Ect), surface registration (Ereg), biological changes (Ebio), and formation of the depth/surface map (Est). This error can be reduced with further information from tool tracking and touching a tool tip to the tissue.
  • the sequences of images 1514 may be stereo image pairs or a sequence of single images from a mono-view.
  • the sequence of images 1514 along with kinematics information may be adaptively fused together to translate a model of the tool tip (M ET ) into the endoscopic camera frame of reference (ECM).
  • the model of the tool tip formed from the adaptive fusion based tool tracking may be used to estimate the tool tip location 1518.
  • An estimation error (E ET ) between the actual tip location 1520 and the estimated tip location 1518 may be made small by the tool tracking methods described herein.
  • One or more of the robotic surgical tools may include a needle at its tip as the end effector.
  • the operational error E O p between the actual location of the tumor and the actual location of tool tips is a sum of errors.
  • a pre- scanned image 1602A (for example, CT images) may be aligned to a camera coordinate system 1605 and then overlaid as an overlaid image 1602B onto a surface map or depth map 1600 of the tissue surface.
  • the tool 1610L and the second robotic instrument 1610R may be adaptively fused together to more accurately determine the position of the tools within the surgical site.
  • the tool With both position of the overlaid image 1602B and the tools 161OL, 161OR aligned to the camera coordinate system 1605, the tool may be automatically or manually guided to perform a procedure on the tissue.
  • the overlaid image 1602B may represent a tumor on the tissue surface or buried below the tissue surface that would otherwise be hidden from view.
  • an image 1610B of the portion of the tool below the surface may be synthesized in response to the tool tracking information and the adaptive fusion of kinematics with a priori video information of the tool.
  • the one or more robotic surgical tools 1610L and 1610R may be used to take measurements or determine a surface profile. Tissue in a surgical site may be touched with the tool tip of the robotic tool 1610L at a first point and tissue in the surgical site may be touched with the tool tip of the robotic tool 1610R at a second point. A sequence of images of the surgical site including the robotic surgical tools 1610L and 1610R may be captured. Kinematics information and image information of the robotic tools may be adaptively fused together to accurately determine the tool tip locations at the first and second points. The pose of the first tool tip at the first point and the pose of the second tool tip at the second point may be compared to determine a distance between them. The tool tips may be touching external surfaces of a tumor to determine a diameter of the tumor. Alternatively, the tool tips may be touching external surfaces of an organ to measure the diameter of the organ.
  • a sequence of depth maps including a robotic instrument may be analyzed to determine a location of robotic instruments.
  • Robotic instruments can be located in the camera coordinate frame using any depth map in which the tool can be identified.
  • Kinematics datum provides an approximate location for a robotic instrument that is to be tracked. This is a priori knowledge for the next iteration of the tracking problem.
  • the depth map is analyzed in the environs of the approximate location to locate the robotic instrument. Other information may be employed to improve the a priori knowledge, such as a dynamic model of the robotic instrument, or knowledge of the type of procedure being performed under the camera capturing the images.
  • the locating of the robotic instrument may be performed over any number of depth maps.
  • Kinematics datum provides an approximate location for a robotic instrument that is to be tracked. This is a priori knowledge for the next iteration of the tracking problem.
  • the depth map is analyzed in the environs of the approximate location to locate the robotic instrument.
  • Other information may be employed to improve the a priori knowledge, such as a dynamic model of the robotic instrument, or knowledge of the type of procedure being performed under the camera capturing the images. If the robotic tool is obscured, a current optimal estimate of the location of the surgical instrument may be made (an a posteriori estimate) using the a priori knowledge and the depth map.
  • Instantaneous (re-)correction of the kinematics datum may be computed by adaptively fusing together the available kinematics data, visual information, and/or a priori information.
  • the correction to the current state is used to update the ongoing correction of future kinematics data.
  • the correction is simply made to future data without regard for past corrections.
  • the sequence of corrections is analyzed and an optimal correction based on all available past corrections is computed and used to correct the kinematics data. Analysis by synthesis and appearance learning techniques may be used to improve the correction to the current state.
  • Algorithms that locate the surgical instrument in the depth map and provide the optimal kinematic correction can be further optimized by understanding of the relative variances in corrected kinematic residual error vs. the variances in computed robotic instrument location from the depth map.
  • Kinematics is suspected of initially having a large DC bias, but relatively small variance.
  • a well-design image processing sub-system may have substantially zero DC bias but a relatively large variance.
  • An example of an optimal correction that accommodates these two differing noise processes is one generated by a Kalman filter.
  • an image 1610B of the portion of the tool below the surface may be synthesized.
  • the synthesized image portion 1610B may be included in the tool images 400L,400R as synthesized image portion 400B displayed on the display devices 402L,402R of the stereo viewer 312 illustrated in Figure 4.
  • FIG 5B while tracking a robotic surgical tool, it may exit the field of view 510 of a camera entirely such, as illustrated by the tool 10 IF.
  • the robotic surgical tool may no longer be in the field of view as a result of camera movement over the surgical site, robotic surgical tool movement, or a combination of both.
  • the camera may move away from the position of the robotic surgical tool in a surgical site such that the robotic surgical tool is outside the field of view of the camera.
  • the robotic surgical tool may move away from a position of the camera in a surgical site such that the robotic surgical tool is outside the field of view of the camera. In either case, a surgeon may be left guessing where the robotic surgical tool is outside the field of view unless some indication is provided to him in the stereo viewer 312.
  • compass icons or a compass rose may be displayed in the display devices of the stereo viewer 312 to provide an indication where the robotic surgical tool is located outside the field of view and a direction of tool reentrance into the field of view.
  • compass icons for the directions North 420N, South 420S, East 420E, and West 420W as well as directions in between such as North-East 420NE, North- West 420NW, South-East 420SE, and South-West 420SW may be indicated in the stereo viewer to indicate tool reentrance into the field of view of the camera over a surgical site.
  • a robotic surgical tool is tracked in and out of the field of view of the camera. A determination is made by comparing positions of the camera and a tool whether or not it is outside the field of view of the camera by using the available tool tracking information of the robotic surgical tool. It the robotic surgical tool is out of the field of view, then one of a plurality of compass icons 420 may be displayed in the field of view to show a direction of tool reentrance.
  • Additional information may be conveyed to a surgeon looking in the stereo viewer 312 by somewhat altering the one compass icon 420 in the display that indicates the direction of the tool. For example, by flashing the icon fast or slow may indicate the robotic surgical tool is getting closer or farther away respectively from the field of view. As another example, arrowheads may be added to ends of a bar icon to indicate that the robotic surgical tool is move towards or away from the field of view.
  • the icon indicating direction or reentrance may be colored to indicate movement bringing the tool closer to the field of view (e.g., red for getting warmer) or movement taking the tool further away from the field of view (e.g., green for getting colder).
  • the embodiments of the tool tracking system described herein provide an automatic integrated system that is accurate and reliable by adaptively fusing kinematics and visual information, synthesizing images based on a model and prior poses, and employing sequence matching.
  • a number of elements of the tool tracking system are implemented in software and executed by a computer and its processor, such as computer 151 and its processor 302.
  • the elements of the embodiments of the invention are essentially the code segments to perform the necessary tasks.
  • the program or code segments can be stored in a processor readable medium or transmitted by a computer data signal embodied in a carrier wave over a transmission medium or communication link.
  • the processor readable medium may include any medium that can store or transfer information.
  • Examples of the processor readable medium include an electronic circuit, a semiconductor memory device, a read only memory (ROM), a flash memory, an erasable programmable read only memory (EPROM), a floppy diskette, a CD-ROM, an optical disk, a hard disk, a fiber optic medium, a radio frequency (RF) link, etc.
  • the computer data signal may include any signal that can propagate over a transmission medium such as electronic network channels, optical fibers, air, electromagnetic, RF links, etc.
  • the code segments may be downloaded via computer networks such as the Internet, Intranet, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Surgery (AREA)
  • Robotics (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Molecular Biology (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Veterinary Medicine (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Manipulator (AREA)

Abstract

Selon un mode de réalisation, la présente invention concerne un procédé pour le repérage par un système robotique d'un ou de plusieurs instruments robotiques. Le procédé comprend la génération de cinématique pour l'instrument robotique à l'intérieur d'un champ de vision d'une caméra ; la capture d'information d'image dans le champ de vision de la caméra ; et la fusion adaptative de l'information de cinématique et de l'information d'image pour déterminer une information de pose de l'instrument robotique. L'invention concerne également un système médical robotique avec un sous-système de repérage d'outils. L'invention concerne en outre un procédé pour localiser un instrument robotique dans le champ de vision d'une caméra ; un procédé permettant d'indiquer l'entrée d'outils dans le champ de vision de la caméra ; un procédé de chirurgie guidée par l'image ; et un système de repérage d'outils comportant un ordinateur comprenant un processeur pour exécuter un code de programme lisible par ordinateur et un support utilisable par ordinateur comprenant le code de programme lisible par ordinateur.
PCT/US2008/077611 2007-09-30 2008-09-25 Procédés et systèmes de localisation d'outils et de repérage d'outils d'instruments robotiques dans des systèmes chirurgicaux robotiques WO2009045827A2 (fr)

Applications Claiming Priority (6)

Application Number Priority Date Filing Date Title
US11/865,014 2007-09-30
US11/865,015 US8147503B2 (en) 2007-09-30 2007-09-30 Methods of locating and tracking robotic instruments in robotic surgical systems
US11/865,016 2007-09-30
US11/865,016 US8073528B2 (en) 2007-09-30 2007-09-30 Tool tracking systems, methods and computer products for image guided surgery
US11/865,014 US8108072B2 (en) 2007-09-30 2007-09-30 Methods and systems for robotic instrument tool tracking with adaptive fusion of kinematics information and image information
US11/865,015 2007-09-30

Publications (2)

Publication Number Publication Date
WO2009045827A2 true WO2009045827A2 (fr) 2009-04-09
WO2009045827A3 WO2009045827A3 (fr) 2009-07-23

Family

ID=40342114

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2008/077611 WO2009045827A2 (fr) 2007-09-30 2008-09-25 Procédés et systèmes de localisation d'outils et de repérage d'outils d'instruments robotiques dans des systèmes chirurgicaux robotiques

Country Status (1)

Country Link
WO (1) WO2009045827A2 (fr)

Cited By (36)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2010147766A1 (fr) 2009-06-16 2010-12-23 Intuitive Surgical Operations, Inc. Outil de mesure virtuel pour chirurgie à invasion minimale
WO2010147729A1 (fr) * 2009-06-16 2010-12-23 Intuitive Surgical Operations, Inc. Outil de mesure virtuel pour chirurgie à invasion minimale
GB2472066A (en) * 2009-07-23 2011-01-26 Medi Maton Ltd Device for manipulating and tracking a guide tube with radiopaque markers
WO2011083374A1 (fr) 2010-01-08 2011-07-14 Koninklijke Philips Electronics N.V. Asservissement visuel non étalonné utilisant l'optimisation de vitesse en temps réel
WO2011086431A1 (fr) * 2010-01-13 2011-07-21 Koninklijke Philips Electronics N.V. Enregistrement et navigation basés sur l'intégration d'images pour chirurgie endoscopique
FR2960332A1 (fr) * 2010-05-21 2011-11-25 Gen Electric Procede de traitement d'images radiologiques pour determiner une position 3d d'une aiguille.
WO2011128766A3 (fr) * 2010-04-13 2011-12-22 Picard Frederic Procédés et systèmes pour le repérage d'objets
EP2502558A1 (fr) * 2011-03-22 2012-09-26 KUKA Laboratories GmbH Poste de travail médical
WO2014093824A1 (fr) * 2012-12-14 2014-06-19 The Trustees Of Columbia University In The City Of New York Suivi sans marqueur d'outils chirurgicaux robotisés
WO2014122301A1 (fr) * 2013-02-11 2014-08-14 Neomedz Sàrl Appareil de suivi destiné au suivi d'un objet par rapport à un corps
US8830224B2 (en) 2008-12-31 2014-09-09 Intuitive Surgical Operations, Inc. Efficient 3-D telestration for local robotic proctoring
KR20140115575A (ko) * 2013-03-21 2014-10-01 삼성전자주식회사 수술 로봇 시스템 및 그 제어 방법
WO2015022684A1 (fr) * 2013-08-10 2015-02-19 Needleways Ltd. Affichage de trajectoire d'aiguille médicale
US8971597B2 (en) 2005-05-16 2015-03-03 Intuitive Surgical Operations, Inc. Efficient vision and kinematic data fusion for robotic surgical instruments and other applications
WO2015134391A1 (fr) * 2014-03-03 2015-09-11 University Of Washington Outils d'éclairage virtuel haptique
US9280825B2 (en) 2014-03-10 2016-03-08 Sony Corporation Image processing system with registration mechanism and method of operation thereof
WO2016010737A3 (fr) * 2014-07-15 2016-04-07 Sony Corporation Système chirurgical assisté par ordinateur avec mécanisme d'alignement de position et son procédé de fonctionnement
US9439556B2 (en) 2010-12-10 2016-09-13 Wayne State University Intelligent autonomous camera control for robotics with medical, military, and space applications
US9471142B2 (en) 2011-06-15 2016-10-18 The University Of Washington Methods and systems for haptic rendering and creating virtual fixtures from point clouds
US9477307B2 (en) 2013-01-24 2016-10-25 The University Of Washington Methods and systems for six degree-of-freedom haptic interaction with streaming point data
WO2016200887A1 (fr) * 2015-06-09 2016-12-15 Intuitive Surgical Operations, Inc. Recherches de contenu vidéo dans un contexte médical
US9579797B2 (en) 2014-04-10 2017-02-28 Quanser Consulting Inc. Robotic systems and methods of operating robotic systems
KR20170104554A (ko) * 2015-02-19 2017-09-15 소니 주식회사 해부학적 수술 동안 수술 도구 위치확인을 위한 방법 및 시스템
KR20190012999A (ko) * 2017-07-31 2019-02-11 경북대학교 산학협력단 광학 추적 시스템 및 제어 방법
KR20190018726A (ko) * 2016-07-14 2019-02-25 인튜어티브 서지컬 오퍼레이션즈 인코포레이티드 컴퓨터 보조 원격조작 시스템에서의 2차 기구 제어
US10433763B2 (en) 2013-03-15 2019-10-08 Synaptive Medical (Barbados) Inc. Systems and methods for navigation and simulation of minimally invasive therapy
US10531926B2 (en) 2016-05-23 2020-01-14 Mako Surgical Corp. Systems and methods for identifying and tracking physical objects during a robotic surgical procedure
US10667868B2 (en) 2015-12-31 2020-06-02 Stryker Corporation System and methods for performing surgery on a patient at a target site defined by a virtual object
US10751133B2 (en) 2017-03-31 2020-08-25 Koninklijke Philips N.V. Markerless robot tracking systems, controllers and methods
EP3678582A4 (fr) * 2017-09-05 2021-06-02 Covidien LP Commande de caméra pour systèmes robotiques chirurgicaux
WO2022034183A1 (fr) * 2020-08-12 2022-02-17 Lina Medical International Operations Ag Caméra chirurgicale
CN114235010A (zh) * 2021-12-31 2022-03-25 杭州堃博生物科技有限公司 定位追踪组件的验证方法、装置、电子设备与存储介质
WO2022181919A1 (fr) * 2021-02-26 2022-09-01 (주)휴톰 Dispositif et procédé pour fournir un environnement d'opération basé sur la réalité virtuelle
US11571814B2 (en) 2018-09-13 2023-02-07 The Charles Stark Draper Laboratory, Inc. Determining how to assemble a meal
WO2023072146A1 (fr) * 2021-10-26 2023-05-04 北京智愈医疗科技有限公司 Système d'inspection automatique ultrasonore transluminale, procédé de commande, support de stockage lisible par ordinateur et dispositif électronique
WO2024123888A1 (fr) * 2022-12-06 2024-06-13 Vicarious Surgical Inc. Systèmes et procédés de segmentation d'anatomie et de suivi de structure anatomique

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6231526B1 (en) * 1991-06-13 2001-05-15 International Business Machines Corporation System and method for augmentation of surgery
US20060258938A1 (en) * 2005-05-16 2006-11-16 Intuitive Surgical Inc. Methods and system for performing 3-D tool tracking by fusion of sensor and/or camera derived data during minimally invasive robotic surgery
EP1854425A1 (fr) * 2006-05-11 2007-11-14 BrainLAB AG Localisation spatiale pour appareils médicaux avec mesure de localisation redondante et pondération pour prioriser les mesures
EP1886640A1 (fr) * 2006-08-08 2008-02-13 BrainLAB AG Système et méthode de planification pour l'ajustement d'un implant osseux de forme libre

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6231526B1 (en) * 1991-06-13 2001-05-15 International Business Machines Corporation System and method for augmentation of surgery
US20060258938A1 (en) * 2005-05-16 2006-11-16 Intuitive Surgical Inc. Methods and system for performing 3-D tool tracking by fusion of sensor and/or camera derived data during minimally invasive robotic surgery
EP1854425A1 (fr) * 2006-05-11 2007-11-14 BrainLAB AG Localisation spatiale pour appareils médicaux avec mesure de localisation redondante et pondération pour prioriser les mesures
EP1886640A1 (fr) * 2006-08-08 2008-02-13 BrainLAB AG Système et méthode de planification pour l'ajustement d'un implant osseux de forme libre

Cited By (82)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8971597B2 (en) 2005-05-16 2015-03-03 Intuitive Surgical Operations, Inc. Efficient vision and kinematic data fusion for robotic surgical instruments and other applications
US8830224B2 (en) 2008-12-31 2014-09-09 Intuitive Surgical Operations, Inc. Efficient 3-D telestration for local robotic proctoring
US9402690B2 (en) 2008-12-31 2016-08-02 Intuitive Surgical Operations, Inc. Efficient 3-D telestration for local and remote robotic proctoring
WO2010147766A1 (fr) 2009-06-16 2010-12-23 Intuitive Surgical Operations, Inc. Outil de mesure virtuel pour chirurgie à invasion minimale
WO2010147729A1 (fr) * 2009-06-16 2010-12-23 Intuitive Surgical Operations, Inc. Outil de mesure virtuel pour chirurgie à invasion minimale
KR101720047B1 (ko) * 2009-06-16 2017-03-27 인튜어티브 서지컬 오퍼레이션즈 인코포레이티드 최소 침습 수술을 위한 가상 측정 도구
KR20120040687A (ko) * 2009-06-16 2012-04-27 인튜어티브 서지컬 오퍼레이션즈 인코포레이티드 최소 침습 수술을 위한 가상 측정 도구
CN102458293A (zh) * 2009-06-16 2012-05-16 直观外科手术操作公司 用于微创外科手术的虚拟测量工具
CN102458294A (zh) * 2009-06-16 2012-05-16 直观外科手术操作公司 用于微创外科手术的虚拟测量工具
US9492240B2 (en) 2009-06-16 2016-11-15 Intuitive Surgical Operations, Inc. Virtual measurement tool for minimally invasive surgery
US9155592B2 (en) 2009-06-16 2015-10-13 Intuitive Surgical Operations, Inc. Virtual measurement tool for minimally invasive surgery
GB2472066A (en) * 2009-07-23 2011-01-26 Medi Maton Ltd Device for manipulating and tracking a guide tube with radiopaque markers
CN102791214A (zh) * 2010-01-08 2012-11-21 皇家飞利浦电子股份有限公司 采用实时速度优化的不经校准的视觉伺服
JP2013516264A (ja) * 2010-01-08 2013-05-13 コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ リアルタイム速度最適化を使用した校正不要のビジュアルサーボ
US9205564B2 (en) 2010-01-08 2015-12-08 Koninklijke Philips N.V. Uncalibrated visual servoing using real-time velocity optimization
WO2011083374A1 (fr) 2010-01-08 2011-07-14 Koninklijke Philips Electronics N.V. Asservissement visuel non étalonné utilisant l'optimisation de vitesse en temps réel
US8934003B2 (en) 2010-01-08 2015-01-13 Koninklijkle Philips N.V. Uncalibrated visual servoing using real-time velocity optimization
JP2013517031A (ja) * 2010-01-13 2013-05-16 コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ 内視鏡手術のための画像積分ベースレジストレーション及びナビゲーション
WO2011086431A1 (fr) * 2010-01-13 2011-07-21 Koninklijke Philips Electronics N.V. Enregistrement et navigation basés sur l'intégration d'images pour chirurgie endoscopique
CN102711650A (zh) * 2010-01-13 2012-10-03 皇家飞利浦电子股份有限公司 用于内窥镜手术的基于图像整合的配准和导航
US9095252B2 (en) 2010-01-13 2015-08-04 Koninklijke Philips N.V. Image integration based registration and navigation for endoscopic surgery
US9341704B2 (en) 2010-04-13 2016-05-17 Frederic Picard Methods and systems for object tracking
WO2011128766A3 (fr) * 2010-04-13 2011-12-22 Picard Frederic Procédés et systèmes pour le repérage d'objets
FR2960332A1 (fr) * 2010-05-21 2011-11-25 Gen Electric Procede de traitement d'images radiologiques pour determiner une position 3d d'une aiguille.
US8600138B2 (en) 2010-05-21 2013-12-03 General Electric Company Method for processing radiological images to determine a 3D position of a needle
US9439556B2 (en) 2010-12-10 2016-09-13 Wayne State University Intelligent autonomous camera control for robotics with medical, military, and space applications
US9872651B2 (en) 2011-03-22 2018-01-23 Kuka Roboter Gmbh Medical workstation
EP2502558A1 (fr) * 2011-03-22 2012-09-26 KUKA Laboratories GmbH Poste de travail médical
US9471142B2 (en) 2011-06-15 2016-10-18 The University Of Washington Methods and systems for haptic rendering and creating virtual fixtures from point clouds
WO2014093824A1 (fr) * 2012-12-14 2014-06-19 The Trustees Of Columbia University In The City Of New York Suivi sans marqueur d'outils chirurgicaux robotisés
US9753542B2 (en) 2013-01-24 2017-09-05 University Of Washington Through Its Center For Commercialization Methods and systems for six-degree-of-freedom haptic interaction with streaming point data
US9477307B2 (en) 2013-01-24 2016-10-25 The University Of Washington Methods and systems for six degree-of-freedom haptic interaction with streaming point data
CN105377174A (zh) * 2013-02-11 2016-03-02 尼奥梅德兹有限责任公司 用于相对于身体跟踪对象的跟踪设备
WO2014122301A1 (fr) * 2013-02-11 2014-08-14 Neomedz Sàrl Appareil de suivi destiné au suivi d'un objet par rapport à un corps
US10433763B2 (en) 2013-03-15 2019-10-08 Synaptive Medical (Barbados) Inc. Systems and methods for navigation and simulation of minimally invasive therapy
KR102117273B1 (ko) 2013-03-21 2020-06-01 삼성전자주식회사 수술 로봇 시스템 및 그 제어 방법
KR20140115575A (ko) * 2013-03-21 2014-10-01 삼성전자주식회사 수술 로봇 시스템 및 그 제어 방법
WO2015022684A1 (fr) * 2013-08-10 2015-02-19 Needleways Ltd. Affichage de trajectoire d'aiguille médicale
US10226869B2 (en) 2014-03-03 2019-03-12 University Of Washington Haptic virtual fixture tools
WO2015134391A1 (fr) * 2014-03-03 2015-09-11 University Of Washington Outils d'éclairage virtuel haptique
US9280825B2 (en) 2014-03-10 2016-03-08 Sony Corporation Image processing system with registration mechanism and method of operation thereof
US9579797B2 (en) 2014-04-10 2017-02-28 Quanser Consulting Inc. Robotic systems and methods of operating robotic systems
WO2016010737A3 (fr) * 2014-07-15 2016-04-07 Sony Corporation Système chirurgical assisté par ordinateur avec mécanisme d'alignement de position et son procédé de fonctionnement
KR20170104554A (ko) * 2015-02-19 2017-09-15 소니 주식회사 해부학적 수술 동안 수술 도구 위치확인을 위한 방법 및 시스템
US10147188B2 (en) 2015-02-19 2018-12-04 Sony Corporation Method and system for surgical tool localization during anatomical surgery
EP3242590A4 (fr) * 2015-02-19 2018-07-25 Sony Corporation Procédé et système pour localisation d'outil chirurgical pendant une chirurgie anatomique
US9905000B2 (en) 2015-02-19 2018-02-27 Sony Corporation Method and system for surgical tool localization during anatomical surgery
KR101986443B1 (ko) * 2015-02-19 2019-06-05 소니 주식회사 해부학적 수술 동안 수술 도구 위치확인을 위한 방법 및 시스템
CN107530132A (zh) * 2015-02-19 2018-01-02 索尼公司 在解剖手术期间用于手术工具定位的方法和系统
WO2016200887A1 (fr) * 2015-06-09 2016-12-15 Intuitive Surgical Operations, Inc. Recherches de contenu vidéo dans un contexte médical
US10600510B2 (en) 2015-06-09 2020-03-24 Intuitive Surgical Operations, Inc. Video content searches in a medical context
US11806089B2 (en) 2015-12-31 2023-11-07 Stryker Corporation Merging localization and vision data for robotic control
US11103315B2 (en) 2015-12-31 2021-08-31 Stryker Corporation Systems and methods of merging localization and vision data for object avoidance
US10667868B2 (en) 2015-12-31 2020-06-02 Stryker Corporation System and methods for performing surgery on a patient at a target site defined by a virtual object
US11937881B2 (en) 2016-05-23 2024-03-26 Mako Surgical Corp. Systems and methods for identifying and tracking physical objects during a robotic surgical procedure
US10531926B2 (en) 2016-05-23 2020-01-14 Mako Surgical Corp. Systems and methods for identifying and tracking physical objects during a robotic surgical procedure
KR20190018726A (ko) * 2016-07-14 2019-02-25 인튜어티브 서지컬 오퍼레이션즈 인코포레이티드 컴퓨터 보조 원격조작 시스템에서의 2차 기구 제어
US11672616B2 (en) 2016-07-14 2023-06-13 Intuitive Surgical Operations, Inc. Secondary instrument control in a computer-assisted teleoperated system
KR102482803B1 (ko) 2016-07-14 2022-12-29 인튜어티브 서지컬 오퍼레이션즈 인코포레이티드 컴퓨터 보조 원격조작 시스템에서의 2차 기구 제어
US10751133B2 (en) 2017-03-31 2020-08-25 Koninklijke Philips N.V. Markerless robot tracking systems, controllers and methods
KR20190012999A (ko) * 2017-07-31 2019-02-11 경북대학교 산학협력단 광학 추적 시스템 및 제어 방법
KR102019482B1 (ko) 2017-07-31 2019-09-06 경북대학교 산학협력단 광학 추적 시스템 및 제어 방법
US11612450B2 (en) 2017-09-05 2023-03-28 Covidien Lp Camera control for surgical robotic systems
US11925513B2 (en) 2017-09-05 2024-03-12 Covidien Lp Camera control for surgical robotic systems
EP3678582A4 (fr) * 2017-09-05 2021-06-02 Covidien LP Commande de caméra pour systèmes robotiques chirurgicaux
US11607810B2 (en) 2018-09-13 2023-03-21 The Charles Stark Draper Laboratory, Inc. Adaptor for food-safe, bin-compatible, washable, tool-changer utensils
US11597085B2 (en) 2018-09-13 2023-03-07 The Charles Stark Draper Laboratory, Inc. Locating and attaching interchangeable tools in-situ
US11597084B2 (en) 2018-09-13 2023-03-07 The Charles Stark Draper Laboratory, Inc. Controlling robot torque and velocity based on context
US11597086B2 (en) 2018-09-13 2023-03-07 The Charles Stark Draper Laboratory, Inc. Food-safe, washable interface for exchanging tools
US11597087B2 (en) 2018-09-13 2023-03-07 The Charles Stark Draper Laboratory, Inc. User input or voice modification to robot motion plans
US11872702B2 (en) 2018-09-13 2024-01-16 The Charles Stark Draper Laboratory, Inc. Robot interaction with human co-workers
US11628566B2 (en) 2018-09-13 2023-04-18 The Charles Stark Draper Laboratory, Inc. Manipulating fracturable and deformable materials using articulated manipulators
US11571814B2 (en) 2018-09-13 2023-02-07 The Charles Stark Draper Laboratory, Inc. Determining how to assemble a meal
US11648669B2 (en) 2018-09-13 2023-05-16 The Charles Stark Draper Laboratory, Inc. One-click robot order
US11673268B2 (en) 2018-09-13 2023-06-13 The Charles Stark Draper Laboratory, Inc. Food-safe, washable, thermally-conductive robot cover
WO2022034183A1 (fr) * 2020-08-12 2022-02-17 Lina Medical International Operations Ag Caméra chirurgicale
KR20220122862A (ko) * 2021-02-26 2022-09-05 (주)휴톰 가상 현실 기반의 수술 환경을 제공하는 장치 및 방법
KR102519114B1 (ko) * 2021-02-26 2023-04-07 (주)휴톰 가상 현실 기반의 수술 환경을 제공하는 장치 및 방법
WO2022181919A1 (fr) * 2021-02-26 2022-09-01 (주)휴톰 Dispositif et procédé pour fournir un environnement d'opération basé sur la réalité virtuelle
WO2023072146A1 (fr) * 2021-10-26 2023-05-04 北京智愈医疗科技有限公司 Système d'inspection automatique ultrasonore transluminale, procédé de commande, support de stockage lisible par ordinateur et dispositif électronique
CN114235010A (zh) * 2021-12-31 2022-03-25 杭州堃博生物科技有限公司 定位追踪组件的验证方法、装置、电子设备与存储介质
WO2024123888A1 (fr) * 2022-12-06 2024-06-13 Vicarious Surgical Inc. Systèmes et procédés de segmentation d'anatomie et de suivi de structure anatomique

Also Published As

Publication number Publication date
WO2009045827A3 (fr) 2009-07-23

Similar Documents

Publication Publication Date Title
US8108072B2 (en) Methods and systems for robotic instrument tool tracking with adaptive fusion of kinematics information and image information
US8792963B2 (en) Methods of determining tissue distances using both kinematic robotic tool position information and image-derived position information
US8147503B2 (en) Methods of locating and tracking robotic instruments in robotic surgical systems
WO2009045827A2 (fr) Procédés et systèmes de localisation d'outils et de repérage d'outils d'instruments robotiques dans des systèmes chirurgicaux robotiques
US20230107693A1 (en) Systems and methods for localizing, tracking and/or controlling medical instruments
US11504095B2 (en) Three-dimensional imaging and modeling of ultrasound image data
EP3081184B1 (fr) Système et procédé de navigation à base d'images fusionnées avec marqueur de placement tardif
US20180220883A1 (en) Image-based global registration system and method applicable to bronchoscopy guidance
US8971597B2 (en) Efficient vision and kinematic data fusion for robotic surgical instruments and other applications
US7889905B2 (en) Fast 3D-2D image registration method with application to continuously guided endoscopy
US20180303550A1 (en) Endoscopic View of Invasive Procedures in Narrow Passages
US20130123801A1 (en) Method of real-time tracking of moving/flexible surfaces
WO2010046802A1 (fr) Procédé et système de localisation sur base d'image
Doignon et al. Pose estimation and feature tracking for robot assisted surgery with medical imaging
US20230123621A1 (en) Registering Intra-Operative Images Transformed from Pre-Operative Images of Different Imaging-Modality for Computer Assisted Navigation During Surgery
US10832422B2 (en) Alignment system for liver surgery
Khare et al. Toward image-based global registration for bronchoscopy guidance
Serna-Morales et al. Acquisition of three-dimensional information of brain structures using endoneurosonography
EP3931799B1 (fr) Suivi de dispositif d'intervention
Deguchi et al. Camera motion tracking of real bronchoscope using epipolar geometry analysis and CT-derived bronchoscopic images
Niemann et al. Image-Based Modeling and its Application in Image Processing
Cheng Stereoscopic Motion Tracking in Biomedical Visualisation
Farag et al. Computer Assisted Radiology and Surgery (CARS-2004)

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 08836520

Country of ref document: EP

Kind code of ref document: A2

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 08836520

Country of ref document: EP

Kind code of ref document: A2