Patent Application
20150145953
The present invention relates to an image completion system for an in-image cutoff region, an image processing device, and program therefor, and more particularly, relates to an image completion system for an in-image cutoff region, an image processing device, and a program therefor, which completes image information on an object space imaged in the state where a portion thereof is hidden by a predetermined member, with other image information imaged in another line-of-sight direction.
In recent years, minimally invasive surgery which does not need a large incision and reduces loads imposed on a patient is widespread, and as the minimally invasive surgery, endoscopic surgery is known. The endoscopic surgery is surgery in which a rod-shaped surgical instrument provided with a scalpel, forceps, a puncture needle, or the like on the tip side thereof, and an endoscope are inserted in the body through holes opened at portions on the body surface of a patient, and an operator treats an affected area by manipulating the surgical instrument from the outside of the body of the patient. Such endoscopic surgery includes a mode in which a surgical instrument is directly manipulated by the hands of an operator, as well as a mode assisted by a surgery assistant robot in which a surgical instrument is moved by the operation of a robot arm.
In the above-described endoscopic surgery, however, the operator cannot directly see the affected area and the surrounding area thereof but can visually confirm the affected area with only an endoscopic image on a monitor, and thus the operator is problematically restricted in the visual field. In particular, with regard to an endoscopic image, when a surgical instrument is displayed, an internal space existing on the depth side of the surgical instrument is hidden by the surgical instrument and cannot be visually confirmed. Such a case may be able to be dealt with by the manipulation of, for example, changing the attitude of the endoscope, but the operation during the surgery is troublesome. In addition, the surgical instrument is often close to an affected area during the surgery and is still often displayed somewhere in the endoscopic image even when the attitude of the endoscope is changed. The existence of a cutoff region due to the surgical instrument thus often makes the visual field even narrower, which makes accurate grasping of the space near the affected area further difficult. For this reason, when there is a dangerous site such as a vessel and a nerve which the surgical instruments should not touch, in the cutoff region, the surgical instrument may be unintendedly touch the dangerous site to cause an accident such as bleeding to occur.
Now, Patent Literature 1 discloses a surgery supporting device for processing three-dimensional image data imaged by an MRI (Magnetic Resonance Imaging system) or the like and superimposing the processed three-dimensional image data onto an endoscopic image. This surgery supporting device is configured to extract a specified region in the three-dimensional image to create segmentation image data, subject the segmentation image data to a projecting process to create a surgery assistant image, and superimpose the surgery assistant image onto the endoscopic image.
In addition, Patent Literature 2 discloses an image processing device for establishing correspondences between a stereoscopic endoscope picture imaged during a surgery and a three-dimensional image obtained from image data imaged by an MRI or the like prior to the surgery, and performing registration between the images to compose the images and display the composite image. This image processing device is configured to, when a portion of one of left and right stereoscopic endoscope pictures is cut off by a surgical instrument, geometrically restore feature points of a tissue existing on the back side of the surgical instrument so as to grasp the three-dimensional position of the tissue existing on the back side of the surgical instrument.
In the above-described surgery supporting device of Patent Literature 1, however, it does not mean that, when the surgical instrument is displayed in the endoscopic image, a cutoff region thereof in a depth direction of the endoscopic image is automatically identified to obtain image information on an internal space in the cutoff region, and thus the surgery supporting device cannot solve the above-described problem of the restriction of the visual field of an operator due to the existence of a surgical instrument in an endoscopic image.
In addition, the above-described image processing device of Patent Literature 2 is subject to a condition that a stereoscopic endoscope picture in which a surgical instrument is not displayed is first obtained and the position and the attitude of the stereoscopic endoscope is not changed from those at that time during the surgery, in order to identify, in the stereoscopic endoscope, the three-dimensional position of the tissue on the back side that is hidden by the surgical instrument or the like. It is therefore needed an operation to retract the surgical instruments into a place which is not displayed in the endoscopic picture every time the attitude of the stereoscopic endoscope is changed, which obstructs an smooth operation of the surgery. Furthermore, since the three-dimensional image that has been imaged by an MRI or the like prior to the surgery is superimposed onto the endoscopic picture, if the state of the an internal space imaged in the endoscopic picture changes due to the movement of an organ or the like displayed in the endoscopic picture during the surgery, the correspondences of the same portion cannot be established between the endoscopic picture obtained in real time and the three-dimensional image representing a past state of the internal space having been obtained by the MRI or the like prior to the surgery, and thus the three-dimensional image cannot be superimposed onto the endoscopic picture.
The present invention is devised in light of such problems, and has an object to provide an image completion system for an in-image cutoff region, an image processing device, and a program therefor which, with respect to an image in which a predetermined object space is imaged, if there is a cutoff region where a portion of the image is cut off by a predetermined member, can complete image information on the object space in the cutoff region without troublesome operation even when the condition of the object space changes.
In order to achieve the above-described object, the present invention employs a configuration mainly including a main imaging device for obtaining a main image in which an object space to be monitored is imaged, a completing-purpose imaging device for obtaining a completing image used for completing the main image by imaging the object space in a line-of-sight direction different from that of the main imaging device, a distance measuring device for measuring separating distances between a predetermined reference point and set points at least three of which are set in the object space, a three-dimensional position measuring device for measuring the three-dimensional positions of the main imaging device and the completing-purpose imaging device, an image processing device for completing a portion of the main image with the completing image on the basis of measurement results from the distance measuring device and the three-dimensional position measuring device, wherein the image processing device obtains image information on a cutoff region in the object space that is hidden on the depth side of a member having a known shape by imaging the member in the main image together with the object space, from the completing image, and replaces image information on the member in the main image with the obtained image information or superimposes the obtained image information onto the image information on the member in the main image so as to generate a composite image in which the cutoff region is completed with the completing image.
According to the present invention, a member such as a surgical instrument is imaged in a main image together with an object space, and when image information on a portion of the object space is hidden by the surgical instrument, image information on the depth side of the member in the hidden portion is completed with image information on a real-time completing image, and a composite image that looks as if the member were seen through can be obtained with respect to the main image in real time. As a result, a cutoff region by the member is cancelled by image processing, and the reduction of a visual field in the main image due to the existence of the cutoff region is ameliorated, which allows the visual field to be substantially expanded.
The embodiment according to the present invention will be described below with reference to the drawings.
This image completion system 10 includes imaging device 11 for imaging an image of an object space being an internal space to be monitored formed by an organ K including an affected area to be treated and the surrounding areas thereof, a distance measuring device 12 for measuring separating distances between a predetermined reference point and a large number of set points that are virtually set to objects in the object space, a three-dimensional position measuring device 13 for measuring three-dimensional position of the imaging device 11, and an image processing device 14 for processing the image obtained by the imaging device 11.
The imaging device 11 is configured by a single lens operating field endoscope 16 (main imaging device) for obtaining an operating field image V1 (refer to
As the distance measuring device 12, for example, there are used devices having a well-known structure disclosed in Japanese Patent Laid-Open No. 2010-220787 or the like, and the distance measuring device 12 includes stereo camera 19 that can obtain a stereo image, and distance measuring means 20 that searches for a corresponding point between a pair of stereo images imaged by the stereo camera 19 and calculates distances from the end of the stereo camera 19 to the corresponding point, by a stereo matching method. Note that the descriptions of the structure and the algorithm of the distance measuring device 12 in detail will be omitted since well-known techniques are used therefor, which is not an essential part of the present invention.
Here, the stereo camera 19 is provided integrally with completing-purpose endoscope 17, and is configured so as to be enabled to obtain an almost entire stereo image of a space imaged by the completing-purpose endoscope 17.
In the distance measuring means 20, as schematically shown in
The three-dimensional position measuring device 13 includes markers 22, at least three of which are attached to members to be subjected to position measurement, and a body 23 including light receiving parts 23A for receiving infrared rays emitted by the markers 22. As the three-dimensional position measuring device 13, there are used devices having a well-known configuration which can detect the three-dimensional positions of the markers 22 by tracking the infrared rays following the movements of the markers 22. The description of the structure in detail will be omitted since it is not an essential part of the present invention. Note that as the three-dimensional position measuring device 13, devices making use of various principles or structured can be alternatively used as long as they can detect the three-dimensional positions of the members to be subjected to the position measurement.
Here, the markers 22 are attached to the rear end portion of each surgical instruments S, the operating field endoscope 16, and the completing-purpose endoscope 17, the rear end portions positioned outside the body in the surgery, and the body 23 identifies the sets of three-dimensional coordinates (positions) with respect to the rear end portions, in the reference coordinate system having an origin being a predetermined point. In addition, the sets of three-dimensional coordinates of components that do not move relatively with respect to the rear end portions are calculated from the sets of three-dimensional coordinates of the rear end portions through mathematical operations performed in the body 23 because the surgical instruments S, the operating field endoscope 16, and the completing-purpose endoscope 17 each have a known shape that has been identified in advance. Note that if the operating field endoscope 16 and the completing-purpose endoscope 17 are integrated in such a manner as not to relatively move, the markers 22 may be provided to only one of them. In addition, in the present embodiment, since the completing-purpose endoscope 17 and the stereo camera 19 of the distance measuring device 12 are provided in such a manner as not to relatively move, when the positions of the components of the completing-purpose endoscope 17 are calculated by the three-dimensional position measuring device 13, the positions of the components of the stereo camera 19 are also identified automatically. It is thereby possible to convert the sets of three-dimensional coordinates of the set points P in the stereo camera coordinate system calculated by the distance measuring device 12 into the sets of three-dimensional coordinates in the reference coordinate system on the basis of the measurement result from the three-dimensional position measuring device 13.
Note that if the stereo camera 19 can be relatively move with respect to all the surgical instruments S, the operating field endoscope 16, and the completing-purpose endoscope 17, the markers 22 are attached also to the rear end portion of the stereo camera 19.
The image processing device 14 is configured by a computer formed by a processing unit such as a CPU and a storage such as a memory and a hard drive, and includes a program installed for causing the computer to function as the following means.
This image processing device 14 is configured to obtain, from the completing image V2, image information on cutoff regions in the object space hidden on the depth side thereof by the surgical instruments S displayed in the operating field image V1, and to replace image information on the cutoff regions in the operating field image V1 with the obtained image information or superimposing the obtained image information onto the image information on the cutoff regions in the operating field image V1 so as to perform a process of generating a composite image in which the cutoff regions are completed by the completing image.
Specifically, the image processing device 14 includes set point position identifying means 25 for identifying, with respect to the set points P, sets of three-dimensional coordinates (three-dimensional positions) in the reference coordinate system on the basis of the measurement results from the distance measuring device 12 and the three-dimensional position measuring device 13 and for calculating sets of in-screen coordinates (sets of two-dimensional coordinates) in the screen coordinate system in the operating field image V1 and sets of in-screen coordinates (two-dimensional coordinates) in the screen coordinate system in the completing image V2, completing image transforming means 26 for generating a transformed image V3 (refer to
The procedure of the image completion in the image processing device 14 will be described below.
First, the set point position identifying means 25 converts the sets of three-dimensional coordinates of the set points P in the stereo camera coordinate system calculated by the distance measuring device 12 into the sets of three-dimensional coordinates in the reference coordinate system (refer to
In the above formulae, a set of coordinates (xi, yi, zi) is a set of three-dimensional coordinates of each set point Pi (i=1 to n) in the reference coordinate system. In addition, in the formulae (1) and (2), a set of coordinates (ui, vi) is a set of in-screen coordinates of a set point Pn in the screen coordinate system in the completing image V2, which is a set of two-dimensional coordinates in the horizontal direction in the screen and the vertical direction in the screen. In addition, f is a focal distance of the operating field endoscope, ku is a screen resolution of the completing-purpose endoscope 17 in the horizontal direction in the screen, kV is a screen resolution of the completing-purpose endoscope 17 in the vertical direction in the screen, and a set of coordinates (u0, v0) is a set of coordinates of a point in the horizontal direction in the screen and the vertical direction in the screen, at which the optical axis crosses the image surface of the completing image V2. Here, f, ku, kV, u0, and v0 are constants that have been specified in accordance with the specification or the state of disposition of the completing-purpose endoscope 17, and stored in advance.
Next, the sets of coordinates (xi, yi, zi) of the set points Pi in the reference coordinate system are converted into sets of three-dimensional coordinates (x′i, y′i, z′i) having a reference being a predetermined position of the operating field endoscope 16, on the basis of a relative position relationship between the operating field endoscope 16 and the completing-purpose endoscope 17 based on the measurement result from the three-dimensional position measuring device 13, and further converted into set of in-screen coordinates (u′i, v′i) of the set point Pi in the operating field endoscope 16 by formulae similar to the above formulae (1) and (2).
Next, in the completing image transforming means 26, on the basis of the sets of in-screen coordinates (u′i, v′i) of the set points Pi in the operating field image V1 and the sets of in-screen coordinates (ui, vi) of the set points Pi in the completing image V2, pieces of image information on the points in the completing image V2 are moved, in the completing image V2, to positions corresponding to the sets of in-screen coordinates in the operating field image V1 at which the same portions of the points in the completing image V2 are displayed, whereby the transformed image V3 for the completing image V2 is generated.
In other words here, first, as shown in
First, a virtual region T that has a certain range smaller than the entire completing image V2 is set, and the set points Pi existing in the virtual region T are identified around the unset point Pp. In the example in
Next, the following weight coefficients Wi are calculated in such a manner as to correspond to the set points Pi existing in the virtual region T. Specifically, a separating distance with respect to the unset point Pp is calculated for each set point Pi existing in the virtual region T, and the weight coefficient Wi is calculated from the separating distance using a preset arithmetic formula. These weight coefficients Wi are set so as to be in inverse proportion to the separating distances.
Next, movement vectors T(up, vp) for the unset points Pp are calculated by the following formula, respectively. Note that, here, a number N of set points Pi existing in the virtual region T are defined as set points PTj (j=1, 2, . . . , N), the movement vectors in the completing image V2 that are identified with respect to the set points PTj by the above-described procedure are defined as T(uj, vj), and the above-described weight coefficients corresponding to the separating distances from the unset points Pp are defined as Wj.
The pieces of image information on the unset points Pp in the completing image V2 are thereafter moved in the screen of the completing image V2 according to the amount and the direction of the movement based on the calculated movement vectors T(up, vp). As a result, the transformed image V3 is generated in such a manner that the pieces of image information on the set points Pi and the unset points Pp in the completing image V2 are moved in the same screen so as to convert the completing image V2 into that in the line-of-sight direction of the operating field endoscope 16.
Note that, in the completing image transforming means 26, the movement vectors T(up, vp) of the pieces of image information on the unset points Pp are calculated by the weighted average, but the movement vectors T(up, vp) may be calculated by other methods such as B-spline interpolation on the basis of pieces of position information on the set points Pi.
Next, in the cutoff region identifying means 27, the positions of the cutoff regions occupied by the main body parts S2 in the operating field image V1 are identified as follows. That is, the three-dimensional position measuring device 13 calculates the sets of three-dimensional coordinates of the parts of the surgical instruments S1 in the reference coordinate system. These sets of three-dimensional coordinates are then converted into the sets of in-screen coordinates (two-dimensional coordinates) in the screen coordinate system of the operating field image V1 using arithmetic formulae similar to those in the description of the set point position identifying means 25, and the positions of the cutoff regions in the operating field image V1 are identified. Note that, the identification of the cutoff regions is not limited to the above-described method, and well-known methods may be used in which predetermined colors are applied to the main body parts S2 and the pieces of image information on the operating field image V1 are distinguished on the basis of the colors to identify the cutoff regions.
Thereafter, in the composite image generating means 28, a composite image is generated by performing the following mask process. That is, first, as shown in
The composite image is an image having the operating field image V1 as a base, in which the pieces of image information on the depth sides of the main body parts S2 are completed by the completing image V2 from the completing-purpose endoscope 17 as if the main body parts S2 of the surgical instruments S displayed in the operating field image V1 are made transparent or translucent. Therefore, in the composite image, only the treating parts S1 being the tips of the surgical instruments S imaged in the operating field image V1 are left, and the internal space except for the treating parts S1 that an operator needs during the surgery can be imaged in the operating field image V1, which allows the operating field of the endoscopic image to be substantially expanded.
Note that, in the above-described embodiment, there has been illustrated and described the image completion system 10 for performing image processing to the endoscopic image in endoscopic surgery, but the present invention is not limited to this, and can be applied to image processing to an endoscopic image from a surgery assistant robot for assisting endoscopic surgery, as well as can be applied to, for example, image processing for performing a remote control of a robot arm while obtaining an image from an imaging device such as a camera in an operation in a working space such as a reactor of a nuclear power plant that a human cannot enter and directly see. In this case, the replacement of the above-described surgical instrument S with a member such as a robot arm that has been specified in advance and the application of an algorithm similar to the above make it possible to implement an image completion system that meets the use.
In addition, the configuration of each part of the device in the present invention is not limited to the illustrated exemplary configurations, and can be subjected to various modifications as long as it exhibits substantially similar effects.
The present invention is industrially applicable as a system for completing a restricted visual field by using an imaging device for obtaining an image of the inside of a space that a human cannot directly see.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2012-061285 | Mar 2012 | JP | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/JP2013/057392 | 3/15/2013 | WO | 00 |
