Method for registering a three-dimensional representation of an object to the object itself and device for aiding navigation within an object

Abstract

This registration process involves recording (100, 126) a three-dimensional representation of an object (12), extracting (104-116, 130) a surface of the object from this representation, and performing surface registration (120, 122, 132) of this representation onto the object. The surface registration process includes recording (120) the locations of surface points (Ps) on the object's surface in a reference frame linked to a referencing ancillary (24), and calculating (122, 132) the surface registration from the extracted object surface and this recording (120) of surface point locations. Surface extraction includes a preliminary processing (104, 130) to distinguish at least one internal and one external surface of the object, a selection (106-114, 130) of at least one viewpoint external to the object, and a deletion (116, 130) of said at least one internal surface on a criterion of invisibility from each selected viewpoint.Figure for the abridged version: Fig. 7.

Description

Title of the invention: Method for realigning a three-dimensional representation of an object on the object itself and device for assisting navigation in an object

[0001] The present invention relates to a method for three-dimensional registration of a three-dimensional representation of an object on the object itself, using a referencing ancillary secured to the object without a degree of freedom and a pointer whose tip is locatable in a three-dimensional reference frame linked to the referencing ancillary. It also relates to a computer program comprising instructions for executing the steps of such a method, as well as a device for assisting navigation in an object implementing this registration method.

[0002] The invention applies more particularly to a method of this type comprising: - a recording in memory of a three-dimensional representation of the object; - an extraction, by a computing unit, of a surface of the object in the three-dimensional representation; and - a three-dimensional surface alignment of this three-dimensional representation on the object comprising: • a recording, using the tip of the pointer, of locations of a plurality of surface points on the surface of the object in the three-dimensional reference system linked to the referencing ancillary, and • a calculation, by the calculation unit, of the three-dimensional surface registration from the surface of the object extracted in the three-dimensional representation and the recording of locations of surface points.

[0003] Such a three-dimensional registration method is for example implemented in a device for assisting navigation in an object using an instrument with a functional end such as the surgical navigation device marketed under the registered trademark “Collin Navigation Solutions” and intended primarily for interventions in otorhinolaryngology, but also incidentally in neurosurgery or at the border of these two surgical disciplines. It generally comprises a preliminary three-dimensional angular registration carried out using reference points of the object to facilitate and accelerate the surface registration. The object is then the head of a human or animal patient inside which an intervention, using a surgical instrument with any functional end, must be envisaged by the auditory, nasal, oral, neurosurgical or lateral route. This is an example of an application in- non-limiting industrial aspect of the present invention.

[0004] In rhinology, for example, the referencing ancillary is preferably a headband or a pad fixed stably on the forehead of a patient, human or animal. The reference points generally chosen are a location point of the nasion, a location point of the nasolabial junction and a location point of the left and / or right external orbital rim. They conventionally indicate the main north, south and west and / or east orientations on the patient's face. On the basis of a good angular registration carried out by a practitioner, the known surface registration methods make it possible to obtain a completely satisfactory registration of the three-dimensional representation on the patient's head.

[0005] But in otology, the dimensions of the object of study or intervention are much smaller and its surface generally more complex. In particular, the bony part of a human or animal ear is made up of alternating soft tissues and complex bony walls, in particular dense, aerated or porous bony walls, alveoli, which hinder simple segmentation and make any surface registration difficult to achieve. As a result, even when a good angular registration is previously carried out by a practitioner, the surface registration methods known in otorhinolaryngology do not allow for a satisfactory final registration of the three-dimensional representation on the otological area of ​​interest of the patient's head.

[0006] In other applications where the dimensions of the object of study are also reduced and its surface complex, even outside medical applications, the result is also unsatisfactory.

[0007] It may thus be desirable to provide a three-dimensional registration method which makes it possible to overcome at least some of the aforementioned problems and constraints.

[0008] A method is therefore proposed for three-dimensional recalibration of a three-dimensional representation of an object on the object itself, using a referencing ancillary secured to the object without a degree of freedom and a pointer whose tip can be located in a three-dimensional reference frame linked to the referencing ancillary, comprising: - a recording in memory of a three-dimensional representation of the object; - an extraction, by a computing unit, of a surface of the object in the three-dimensional representation; and - a three-dimensional surface alignment of this three-dimensional representation on the object comprising: • a recording, using the tip of the pointer, of locations of a plurality of surface points on the surface of the object in the three-dimensional reference system linked to the referencing ancillary, and • a calculation, by the calculation unit, of the three-dimensional registration surface from the object surface extracted in the three-dimensional representation and the recording of surface point locations; in which the surface extraction comprises: - preliminary processing of the three-dimensional representation to distinguish at least one internal surface and one external surface of the object; - a selection of at least one point of view external to the object in the three-dimensional representation; and - a deletion of said at least one internal surface in the three-dimensional representation on a criterion of invisibility of this at least one internal surface from each selected point of view.

[0009] This clever specific processing carried out on the three-dimensional representation of the object of study makes it possible to simplify the representation of its surface by removing any internal surface, including any more or less complex bone or alveolar structure that it includes, likely to disturb the surface registration. This results in satisfactory registration even under difficult conditions such as those relating to three-dimensional imaging in otology. This also results in improved registration under less difficult conditions such as those relating to rhinology or laryngology. Other non-medical applications can also benefit from this clever specific processing.

[0010] Optionally: - the pre-processing includes thresholding, for example binarization, of the three-dimensional representation on the basis of a predetermined gray level threshold value, which can be determined automatically or is adjustable; and / or - the deletion of said at least one internal surface comprises obtaining a rendering of the object in the three-dimensional representation previously processed from each selected viewpoint, then a selection of the visible parts of this rendering of the object from each selected viewpoint by a color buffer selection method.

[0011] Also optionally, a recalibration method according to the invention may further comprise: - a memory recording of locations of at least three reference points of the object in the three-dimensional representation; - a three-dimensional angular recalibration, by the calculation unit, of the three-dimensional representation on the object from a recording, using the tip of the stylus, of locations of the reference points of the object in the three-dimensional reference system linked to the referencing ancillary.

[0012] Also optionally, the selection of said at least one external point of view the object includes: - the determination of at least one axis carrying at least one point of view from the reference points and / or at least one remarkable point of the object; - the automatic selection of said at least one point of view defined as belonging to this at least one supporting axis and external to the object.

[0013] Also optionally, the selection of said at least one point of view external to the object comprises: - the determination of a plane geometric figure representative of the reference points; - the determination of an axis orthogonal to this plane geometric figure passing through a remarkable point, for example an isobarycenter, of this plane geometric figure; - the automatic selection of at least a first point of view defined as belonging to this orthogonal axis, external to the object and located at least a first predefined distance from the remarkable point; - the definition of at least one reference surface from said at least one first point of view; and - the automatic selection of other viewpoints defined as belonging to this at least one reference surface and located at least a second predefined distance from said at least one first viewpoint in at least one direction.

[0014] Also optionally: - three reference points are recorded and the plane geometric figure representative of these three reference points is the triangle that they form; - the remarkable point is one of the elements of the set consisting of the center of gravity of this triangle, the center of the circle inscribed in this triangle, the center of the circle circumscribed to this triangle and the orthocenter of this triangle; - a single first point of view is defined, located at a single first predefined distance from the remarkable point; - this first predefined distance is either a constant or a variable proportional to the surface of this triangle or to at least one dimension of the object; - the reference surface is a plane or a three-dimensional conical surface, of the ellipsoid, paraboloid, hyperboloid type, passing through the first point of view; - the second predefined distance is either a constant or a variable proportional to the area of ​​this triangle or to at least one dimension of the object; and - two other viewpoints, respectively four other viewpoints, are selected from the first viewpoint in one direction, respectively two different directions.

[0015] Also optionally, the three-dimensional surface registration of the three-dimensional representation on the object comprises the recording of the locations of several hundred surface points on the surface of the object in the three-dimensional reference system linked to the referencing ancillary, preferably between 200 and 1000 surface points, more preferably between 500 and 600 surface points.

[0016] Also optionally, the object is a bony object, in particular a portion of temporal bone of a human or animal being.

[0017] There is also provided a computer program downloadable from a communication network and / or recorded on a computer-readable medium and / or executable by a processor, comprising instructions for executing the steps of a method for three-dimensional registration of a three-dimensional representation of an object on the object itself according to the invention, when said program is executed by a calculation unit of a device for assisting navigation in the object using an instrument with a functional end, and when this device further comprises: - a memory for storing a three-dimensional representation of the object; - a referencing ancillary intended to be made integral with the object without any degree of freedom; and - a pointer, which can be manipulated by an operator, one tip of which can be located in a three-dimensional reference system linked to the referencing ancillary.

[0018] There is also provided a device for assisting navigation in an object using an instrument with a functional end, comprising: - a memory for storing a three-dimensional representation of the object; - a computing unit designed for three-dimensional recalibration of this three-dimensional representation of the object on the object itself; - a referencing ancillary intended to be made integral with the object without any degree of freedom; - a pointer whose tip can be located in a three-dimensional reference system linked to the referencing ancillary; wherein the computing unit is more specifically designed to perform: - an extraction of a surface of the object in the three-dimensional representation; and - a three-dimensional surface recalibration of this three-dimensional representation- sional on the object comprising: • a recording, by interaction with the tip of the pointer when it is manipulated by an operator, of locations of a plurality of surface points on the surface of the object in the three-dimensional reference system linked to the referencing ancillary, and • a calculation of the three-dimensional surface registration from the surface of the object extracted in the three-dimensional representation and the recording of locations of surface points; and wherein the computing unit is further designed so that the surface extraction comprises: - preliminary processing of the three-dimensional representation to distinguish at least one internal surface and one external surface of the object; - a selection of at least one point of view external to the object in the three-dimensional representation; and - a deletion of said at least one internal surface in the three-dimensional representation on a criterion of invisibility of this at least one internal surface from each selected point of view.

[0019] The invention will be better understood with the aid of the following description, given solely by way of example and with reference to the appended drawings in which: - [Fig.l] schematically represents the general structure of a device for assisting navigation in an object, according to one embodiment of the invention, - [Fig.2] more precisely represents an example of functional configuration of an object surface extraction software module of the navigation aid device of [Fig.l]; - [Fig.3A] illustrates an example of a plane section of the temporal bone portion of a human being; - [Fig.3B] represents the planar section of the temporal bone part of [Fig.3A] after application of a preliminary treatment by execution of the software module of [Fig.2]; - [Fig.4] schematically illustrates the application of a principle of selection of at least one external point of view to an object by execution of the software module of [Fig.2], according to an embodiment of the invention; - [Fig.5] schematically illustrates the application of a principle of reconstitution of an external surface of an object by execution of the software module of [Fig.2]; - [Fig.6] illustrates different views of a three-dimensional representation of an object before and after removal of at least one internal surface of this object by executing the software module of [Fig.2]; and - [Fig.7] illustrates the successive stages of a three-way recalibration process dimensional representation of a three-dimensional representation of the object of [Fig.l] on itself, as it can be implemented by the navigation aid device of [Fig.l].

[0020] The device 10 for assisting navigation in an object 12 shown schematically in [Fig. 1] comprises a movable navigation unit 14 comprising a main box 16 with electrical, electronic and software components, for example mounted on casters and provided with a grip handle (not shown) to be moved by a practitioner, and a display screen 18, for example a touch screen, secured to the main box 16. In the non-limiting example illustrated in [Fig. 1], it is more precisely a surgical navigation device intended mainly for interventions in otorhinolaryngology (ENT), but also usable in neurosurgery or at the border between these two surgical disciplines (i.e. in otoneurosurgery for example), such as the aforementioned “Collin Navigation Solutions” surgical navigation device.Object 12 is then the head of a human or animal patient inside which an intervention must be considered by auditory, nasal, oral or even oto- and / or neuro-surgical route when it is a lateral surgery of the base of the skull.

[0021] To do this, the navigation aid device 10 further comprises an instrument 20 with any functional end 22, such as an ENT or neurosurgical instrument with a suction cannula, rotating tip, forceps, simple point, or any other functional end 22 useful for an ENT surgery and / or (oto)neurosurgery intervention via the auditory, nasal, oral, neurosurgical or lateral route.

[0022] For the implementation of a method for three-dimensional resetting of a three-dimensional representation of the object 12 on the object 12 itself, as well as a method of navigation assisted by tracking, on the display screen 18, a movement of the functional end 22 of the surgical instrument 20 in interior cavities of the head 12 of the patient, the navigation aid device 10 further comprises: - a referencing ancillary 24 intended to be fixed, directly or indirectly, on the head 12 of the patient to be integral therewith without any degree of freedom, and - a resetting pointer 28, a distal tip 30 of which can be located in a three-dimensional reference frame linked to the referencing ancillary 24, this reference frame therefore also being linked to the head 12 of the patient when the referencing ancillary 24 is fixed there.

[0023] The methods of fixing the referencing ancillary 24 on the head 12 are known per se and will not be detailed. They are, for example, advantageously defined in such a way as to ensure stability of the fixation capable of inducing a resetting error, and therefore a submillimetric navigation tracking error.

[0024] According to a simple and advantageous known practice illustrated in a non-limiting manner in [Fig. 1], the referencing ancillary 24 comprises at least one electromagnetic or optical location element 26 and the resetting pointer 28 comprises at least one corresponding electromagnetic or optical element capable of interacting directly with the location element 26 of the referencing ancillary 24 or, as a variant, indirectly via another electromagnetic or optical location element remote from the object 12 and / or the resetting pointer 28. Whatever the technology and arrangement chosen, the essential thing is to have a referencing ancillary 24 secured to the object 12 without any degree of freedom, thus fulfilling its function as a reference frame linked to the object 12, and a resetting pointer 28 whose distal tip 30 can be located in this reference frame by direct or indirect interaction.

[0025] [Fig.l] illustrates more precisely a referencing ancillary 24 taking the form of an electromagnetically emitting pellet screwed laterally into the patient's skull in direct contact with the bony surface of the petrous bone or the mastoid part of the temporal bone of one of his ears, for an otological intervention for example by suprapetrous, translabyrinthine, retrolabyrinthine, retrosigmoid route, or any other route suitable for an intervention in lateral surgery, in particular in lateral surgery of the base of the skull. In this case, an incision is made by the practitioner and the skin is locally folded or pushed back to place and fix the pellet in contact with the bone. In this case also, the area of ​​interest is of reduced size, generally much smaller than the object of study in rhinology or laryngology. Similarly, the bone surface is particularly complex and difficult to grasp for three-dimensional registration.

[0026] But any other well-known referencing ancillary is also suitable, such as: - a forehead strip fixed stably on the patient's forehead or a forehead pad fixed stably by double-sided adhesive tape on the patient's forehead, in particular for rhinology surgery; - an oral device placed in the patient's mouth between their teeth, particularly for laryngology surgery; or - any other form of reference ancillary more generally suitable for ENT surgery or (oto)neurosurgery.

[0027] [Fig.l] also illustrates more precisely a resetting pointer 28 taking the form of a stylus, for example a push-button stylus. This configuration is particularly advantageous for easily collecting surface points. But other configurations are possible and known, such as, for example, the use of any tip associated with a pedal or with a voice or touch control of the device 10. More generally, any device making it possible to point precisely and quickly at a plurality of surface points and to easily control the collection of their locations is suitable.

[0028] It will be noted that the way in which the distal tip 30 of the resetting stylet 28 can be located in the three-dimensional reference frame linked to the referencing ancillary 24 (using an optical or electromagnetic sensor with a receiving function for example) is well known, so that the electromagnetic or optical technology implemented for such localization will not be detailed. And it is according to this same technology that the surgical instrument 20 can also be located (also using an optical or electromagnetic sensor with a receiving function for example) in the three-dimensional reference frame linked to the referencing ancillary 24 (also using an optical or electromagnetic sensor with a transmitting function for example).More precisely, in order for its functional end 22, potentially interchangeable, to be itself directly localizable in the three-dimensional reference frame linked to the referencing ancillary 24, a calibration of the surgical instrument 20 may be necessary. This calibration, carried out using a calibration plot for example integral with the referencing ancillary 24, or using the resetting stylet 28, is also well known, so that it will not be detailed either.

[0029] It will just be noted that in the case of an application in otology as illustrated in [Fig.l], the distal tip 30 of the resetting stylet 28 is advantageously very fine, in particular in the form of a tip whose end is of submillimeter radius, rather than a ball end of millimeter radius (i.e. generally about 1.5 mm) as is generally the case in rhinology where the resetting is done on the skin. It must indeed advantageously be able to be directly in contact with the rock or the mastoid part of the temporal bone to grasp its complex external surface very precisely without being traumatic.It thus optionally but advantageously has the following characteristics: it can be screwed onto a base of the recalibration stylus 28 and of a height such that its end coincides with the calibrated virtual point of the stylus (for example 11 mm + / - 0.05 mm); its base is machined to be compatible with the base of the stylus (for example according to a metric M3 thread); its material is resistant to autoclaving at at least 134 °C for at least 18 minutes and is biocompatible, for example polyetheretherketone or PEEK (from the English "PolyEtherEtherKetone"); the radius of its end is 0.3 mm + / - 10%.

[0030] It will also be noted that during an intervention implementing the three-dimensional recalibration of a three-dimensional representation of the object 12 on the object 12 itself, the calibration of the surgical instrument 20 or the aforementioned assisted navigation, it It is appropriate to isolate the head 12 of the patient on which the referencing ancillary 24, the re-alignment stylet 28 and the surgical instrument 20 are fixed from any disturbance, in particular from any disturbance due to too great proximity to the movable navigation unit 14. This is the reason why the operating field ZI in which the head 12 of the patient, the referencing ancillary 24, the re-alignment stylet 28 and the surgical instrument 20 are located is distant and well isolated from the navigation aid field Z2 in which the movable navigation unit 14 is located. This distance, or this isolation, is marked in [Fig.l] by the horizontal line T in broken lines. It is particularly relevant when referencing is carried out by electromagnetic technology, which is, by way of non-limiting example, assumed to be the case of the installation of [Fig.l].

[0031] To interact with the referencing ancillary 24, the resetting stylet 28 and the surgical instrument 20, the main box 16 is provided with a spatial localization interface 32 with electromagnetic field creation. This is an electronic element making it possible to determine the three-dimensional spatial coordinates of a specific sensor, such as those embedded in the referencing ancillary 24, the resetting stylet 28 and the surgical instrument 20. Such a spatial localization interface 32 is well known, so it will not be detailed.

[0032] The main box 16 is further provided with an interface 34 for connection to other possible peripherals: in particular a keyboard 36, a mouse 38 and / or a remote control 40. The remote control 40 is for example connected by an infrared type IR wireless link, which allows it to be used in the operating field ZI. The interface 34 may further comprise a USB port, a CD-ROM reader, or any other video input making it possible to record a three-dimensional representation of the object 12 previously acquired according to any three-dimensional imaging mode.

[0033] The main box 16 further comprises several functional modules which will be described below. In the example described, these modules are of a software nature. Thus, the main box 16 comprises a computer-type element 42 comprising a processing unit 44, interacting with the spatial location interface 32, the interface 34 and the display screen 18, and an associated memory area 46 in which a data storage area 48 and several computer programs 50 to 62 or several functions of the same computer program are recorded. These computer programs comprise instructions designed to be executed by the processing unit 44 in order to carry out the functions of the software modules. They are presented as distinct, but this distinction is purely functional. They could just as well be grouped according to all possible combinations into one or more software programs.Their functions could also be at least partly micro. programmed or micro-wired in dedicated integrated circuits, such as digital circuits. Thus, as a variant, the computer 42 could be replaced by an electronic device composed solely of digital circuits (without a computer program) for carrying out the same functions. Also as a variant, the aforementioned computer programs could be remote and accessible by the computer 42 via the Internet. Generally speaking, even if all the aforementioned software and memory components are presented as being gathered in the same main box 16, they could just as well be dispersed in separate hardware elements, or even distant from each other, but interconnected in a network (data transmission bus, local network, wide area network, Internet, etc.).

[0034] The main box 16 thus firstly comprises a software module 50, intended to be executed by the processing unit 44, for recording in memory 46, more precisely in the data storage area 48 (whatever the nature of this data storage area: RAM, ROM, volatile, flash, virtual, etc.), at least one three-dimensional representation of the object 12, otherwise generally called a “series”, according to at least one three-dimensional imaging mode, preferably according to one or more imaging modes by scanner (i.e. by computed tomography) or MRI (i.e. for “Magnetic Resonance Imaging”) with or without injection of contrast product(s). Generally speaking, any three-dimensional imaging mode allowing extraction of a surface of the object 12 is suitable, knowing that the series obtained by scanner are mainly used in ENT surgery.

[0035] For an application in otology, it will be noted that high definition imaging may prove necessary given the small dimensions and surface complexities of the intervention area. For example, one could opt for scanner or MRI imaging with an increment of 0.1 to 0.5 mm and a thickness of 0.1 to 0.5 mm.

[0036] The main box 16 further comprises a software module 52, intended to be executed by the processing unit 44, for extracting a surface of the object 12 in each three-dimensional representation of the object 12 recorded in the data storage area 48. A non-limiting example of functional configuration of this surface extraction software module 52 will be detailed with reference to [Fig. 2]. This can generally be done automatically if good known conditions for acquiring each three-dimensional representation have been respected.

[0037] The main box 16 further comprises a software module 54, intended to be executed by the processing unit 44, for three-dimensional angular recalibration of each three-dimensional representation of the object 12 recorded in the data storage area 48 on the object 12 itself.

[0038] According to an example of an angular resetting protocol advantageously implemented but not limiting, in particular if it is a question of angularly resetting a first representation three-dimensional representation, it is appropriate to locate at least three predetermined anatomical reference points of the object 12 in the three-dimensional representation concerned and to record them in the data storage area 48, then to locate these same reference points in the three-dimensional reference system linked to the referencing ancillary 24 using the tip 30 of the stylus 28 in a manner known per se and to also record them in the data storage area 48.If it is a question of angularly realigning a second or subsequent three-dimensional representation, while the first three-dimensional representation has already been angularly realigned, it is sufficient to locate the reference points of the object 12 in the three-dimensional representation concerned and to record them in the data storage area 48 without needing to locate these same reference points again in the three-dimensional reference system linked to the referencing ancillary 24 using the stylus 28, since the angular realignment can advantageously take advantage of the locations recorded using the stylus 28 for the first three-dimensional representation given that they have no reason a priori to be different.The angular recalibration is then carried out in a manner known per se by a two-by-two matching of the reference points respectively located in the three-dimensional representation concerned and in the three-dimensional reference system linked to the referencing ancillary 24.

[0039] Reference points should be chosen that are as little aligned as possible to make the angular resetting effective and precise. For example, when the referencing ancillary 24 is a pellet screwed laterally into the patient's skull in direct contact with the bony surface of the petrous bone or the mastoid part of the temporal bone of one of his ears, for an otological intervention, for example by suprapetrous, translabyrinthine, retrolabyrinthine, retrosigmoid route, or any other route suitable for a lateral surgery intervention, the reference points generally chosen are distributed around the area of ​​interest or operating area quite far apart in the temporal bone. They can, by convention, indicate main north, south, west and / or east orientations and are generally three in number.Any other reference point can be freely chosen depending on the location where the referencing ancillary 24 is fixed, the intended application, and even more generally the nature of the object 12.

[0040] In otology at least, given that the location of the reference points in the three-dimensional reference system linked to the referencing ancillary 24 is done using the tip 30 of the stylus 28 in direct contact with the temporal bone, it may be advantageous to locate the reference points on the external bone surface of the object 12 as extracted in the three-dimensional representation by execution of the software module 52 for better precision of the angular resetting.

[0041] The main box 16 further comprises a software module 56, intended to be executed by the processing unit 44, of three-dimensional surface registration of each three-dimensional representation of the object 12 recorded in the data storage area 48 on the object 12 itself.

[0042] According to an example of a surface registration protocol advantageously implemented but not limiting, if it is a question of registering a first three-dimensional representation, it is first of all necessary to locate a plurality of surface points, in particular other than the reference points, on the surface of the object 12 in the three-dimensional reference frame linked to the referencing ancillary 24 using the tip 30 of the stylus 28 and to record them in the data storage area 48. It is then a question of calculating the three-dimensional surface registration from this recording of locations of surface points and the surface of the object 12 as extracted by the software module 52.

[0043] More precisely, the three-dimensional surface registration of a three-dimensional representation on the object 12 comprises the calculation of a transition matrix between the locations of the surface points recorded previously and their locations on the surface of the object 12 as extracted in the three-dimensional representation. The prior angular registration makes it possible to facilitate the calculation of this transition matrix by limiting the possible solutions.

[0044] If it is a question of recalibrating a second or subsequent three-dimensional representation, while a first three-dimensional representation has already been recalibrated according to the aforementioned surface recalibration protocol, it is sufficient to extract a surface of the object 12 in the three-dimensional representation concerned by executing the software module 52 and to record it in the data storage area 48, then to directly carry out a three-dimensional surface recalibration of this second or subsequent three-dimensional representation from the recording of the locations of surface points carried out for the surface recalibration of the first three-dimensional representation. This way of proceeding is very advantageous because the recording of the locations of surface points can be tedious and time-consuming.

[0045] More precisely, the three-dimensional surface registration of the second or subsequent three-dimensional representation on the object 12 comprises the calculation of a second transition matrix between the locations of the surface points previously recorded for the surface registration of the first three-dimensional representation and their locations on the surface of the object 12 as extracted in the second or subsequent three-dimensional representation. Similarly, the prior angular registration of the second or subsequent three-dimensional representation makes it possible to facilitate the calculation of this second transition matrix by limiting the possible solutions.

[0046] Other known methods, in particular by recalibrating series between them allowing to use the same passage matrix, moreover widely discussed in the scientific literature, are also applicable but will not be detailed.

[0047] Concretely, numerous surface points can be located using the tip 30 of the stylet 28 by moving it over the patient's head 12, in particular over the temporal bone in otology or otoneurosurgery, all around and / or in the vicinity of the reference points. There is no real limit. For example, several hundred surface points are advantageously located in the three-dimensional reference system linked to the referencing ancillary 24 for the most precise surface registration possible in ENT surgery or (oto)neurosurgery, preferably between 200 and 1000 surface points, more preferably between 500 and 600 surface points.

[0048] In accordance with the general principles of the present invention, a functional configuration of the surface extraction software module 52 will now be detailed according to a possible embodiment illustrated in [Fig.2].

[0049] The software module 52 thus comprises functional means 52-1 (i.e., computer program instruction lines) configured for prior processing of each three-dimensional representation of the object 12 recorded in the data storage area 48 to distinguish at least one internal surface and one external surface of the object 12. This prior processing comprises, for example, thresholding of each three-dimensional representation on the basis of a predetermined gray level threshold value, which can be determined automatically or adjustable, knowing that the threshold value can be different from one three-dimensional representation to another.In ENT imaging, and in particular in imaging of the temporal bone part of a human or animal being, each three-dimensional representation generally has gray levels that are quite discriminating between empty parts with very low gray levels, bone parts with high gray levels and soft tissues with intermediate gray levels. It is therefore quite easy, by histogram analysis, to automatically determine the threshold value that is suitable for retaining by thresholding only the bone parts of each three-dimensional representation. These bone parts comprise the external surface of the object 12 as well as alveoli or internal walls, more generally more or less complex and / or alveolar internal structures, which constitute said at least one internal surface of the object 12. Alternatively, the threshold value can be chosen by interaction using the peripherals 36, 38 and / or 40.

[0050] A planar section of a three-dimensional representation of a temporal bone part of a human being is illustrated as an example in [Fig. 3A]. There we can clearly see the three gray levels which distinguish the empty parts 64, the soft tissues 66 and the bone parts 68. By applying a thresholding as defined above, we very easily obtain the representation of [Fig. 3B] which now only includes the bone parts 68. It will be noted that the thresholding carried out by executing the means func tionals 52-1 may consist of simply setting to zero all gray levels below the threshold value. Alternatively, it may consist of binarizing each three-dimensional representation of the object 12 around the threshold value.

[0051] Returning to [Fig.2], the software module 52 further comprises functional means 52-2 (i.e., computer program instruction lines) configured to select at least one viewpoint external to the object 12 in its three-dimensional representation.

[0052] Each viewpoint can be selected by interaction using the peripherals 36, 38 and / or 40. But it is also advantageously possible to define it or them and select it automatically from the reference points chosen to carry out the angular resetting and / or from one or more other remarkable point(s) of the object 12, such as its isobarycenter for example. This automatic selection then comprises for example the determination of at least one axis carrying at least one viewpoint from the reference points REF1, REF2, REF3 and / or another remarkable point of the object 12, as well as the automatic selection of said at least one viewpoint defined as belonging to this at least one carrier axis and external to the object 12.

[0053] Thus, for example, a plane geometric figure representative of the reference points is first determined. With three reference points, this plane geometric figure is preferably the triangle that they form. With more than three reference points, this plane geometric figure is, for example, the polygon that results from an orthogonal projection in the plane minimizing a sum of distances with respect to the reference points. Other geometric figures representative of the reference points and determinable automatically can be envisaged, even non-planar ones.

[0054] An axis orthogonal to the plane geometric figure and passing through a remarkable point, for example an isobarycenter, of this plane geometric figure can also be determined automatically. At least one first viewpoint defined as belonging to this orthogonal axis, external to the object 12 and located at least a first predefined distance from the remarkable point can then be automatically selected. This first distance can be defined as a predetermined constant, as a variable proportional to the surface area of ​​the plane geometric figure, or as a variable proportional to at least one dimension of the object 12 such as its volume, its surface area or its size.

[0055] Other viewpoints can then be automatically selected from the first viewpoint(s) by defining one or more reference surfaces from the first viewpoint(s) and defining the other viewpoints as belonging to this or these reference surfaces and located at least a second distance away. predefined from the first point(s) of view in at least one direction. Each reference surface is for example a plane or a three-dimensional conical surface, of the ellipsoid, paraboloid, hyperboloid type, passing through the or one of the first point(s). The second distance can also be defined as a predetermined constant, as a variable proportional to the surface of the plane geometric figure, or as a variable proportional to at least one dimension of the object 12 such as its volume, its surface or its size. More simply, the second distance can be defined as a ratio of the first distance.

[0056] [Fig.4] illustrates a non-limiting example of automatic selection of viewpoints.

[0057] In this example, three reference points REF1, REF2, REF3 are recorded and the plane geometric figure representative of these three reference points is the triangle T that they form. The remarkable point RQ is one of the elements of the set consisting of the center of gravity of this triangle T, the center of the circle inscribed in this triangle T, the center of the circle circumscribed to this triangle T and the orthocenter of this triangle T.

[0058] A first possible point of view PV1 is defined as belonging to the axis orthogonal to T passing through RQ, external to the object 12 and located at a first predefined distance dl from the remarkable point RQ. According to a first variant, this distance dl is a predetermined constant. According to a second non-limiting variant, it is equal to kS, where k is a strictly positive constant and S is the surface area of ​​the triangle T. According to other variants, it could be proportional to at least one dimension of the object 12 such as its volume, its surface area or its size.

[0059] From the first viewpoint PV1, a rectangle with PV 1 as the center is defined in the plane passing through PV 1 and parallel to the triangle T. Four other possible viewpoints PV2, PV3, PV4, PV5 are then defined as being the four vertices of this rectangle. In other words, they are defined from the first viewpoint PV1 according to two different directions which are those of the two diagonals of the rectangle considered.

[0060] Finally, from the five possible viewpoints PV1, PV2, PV3, PV4, PV5, only a part (for example the four viewpoints PV2, PV3, PV4, PV5) or all five can be selected.

[0061] Another example of automatic selection of viewpoints could be to define the remarkable point RQ as being the isobarycenter of the object 12 as defined in its three-dimensional representation, then to define at least three possible viewpoints as belonging respectively to the axes (RQ, REF1), (RQ, REF2) and (RQ, REF3), external to the object 12 and at certain predefined distances from REF1, REF2 and REF3.

[0062] Returning to [Fig. 2], the software module 52 further comprises functional means 52-3 (i.e., computer program instruction lines) configured for deletion of the internal surface(s) of the object 12 in each three-dimensional representation, on a criterion of invisibility of this or these internal surface(s) from each selected viewpoint, so that only its external surface remains for surface registration.

[0063] According to a preferred embodiment, the deletion of the internal surface(s) of the object 12 in any three-dimensional representation of the object 12 comprises obtaining a rendering of the object 12 in this three-dimensional representation previously processed by thresholding from each selected viewpoint, then a selection of the visible parts of this rendering of the object from each selected viewpoint by a color buffer selection method, such as for example the method defined in the class "vtkHardwareSelector" of the free software library VTK. This image processing, or any other equivalent of deletion of hidden surfaces, is known per se so that it will not be detailed.For each selected viewpoint, it is necessary to follow each direction from the viewpoint towards the object 12 in its three-dimensional representation previously processed by thresholding until encountering a first brightness gradient representative of the external surface, then to delete any subsequent brightness gradient.

[0064] An example of the result of deleting the internal surface(s) of the object 12 in its three-dimensional representation on a criterion of invisibility of this or these internal surface(s) from each selected viewpoint is illustrated in [Fig. 5] for the aforementioned viewpoints PV2, PV3, PV4 and PV5. Each viewpoint PV2, PV3, PV4 or PV5 makes it possible to obtain the reconstruction of only a part of the external surface of the object 12. By multiplying the viewpoints, it is possible to merge the partial reconstructions and obtain a final reconstruction 70 that is as complete as possible of this external surface of the object 12.

[0065] [Fig. 6] thus illustrates in the upper part and from left to right side, top and rear views of a “raw” three-dimensional representation of the mastoid bone portion of a human temporal bone, that is to say after thresholding to remove the soft tissues, but before removing the underlying structures that constitute the surfaces other than the external surface of the object 12. Equivalent side, top and rear views are illustrated from left to right in the lower part of [Fig. 6], but after selection of viewpoints and removal of the internal surface(s) of the object 12 from these viewpoints. A three-dimensional representation of the external surface of the object 12 is clearly simplified and therefore more usable than the “raw” three-dimensional representation for producing a improved surface registration, or even improved angular registration.

[0066] Returning to [Fig.l], the main box 16 further comprises a software module 58, intended to be executed by the processing unit 44, for selecting a display to assist navigation carried out using the surgical instrument 20 when one or more of the three-dimensional representations recorded in the data storage area 48 has been recalibrated by execution of the software modules 52, 54 and 56.

[0067] In particular when two recalibrated three-dimensional representations are available in memory 46, for example one in the form of a volume image of the object 12 by scanner, the other in the form of a volume image of the object 12 by magnetic resonance, the software module 58 manages the possible selection of a simultaneous display of several planar reconstructions in section of the object 12, including the possibility of displaying two planar reconstructions of the same section of the object 12 obtained respectively from one and the other of the first and second three-dimensional representations.

[0068] In ENT surgery, the planar sectional reconstructions of the object 12 generally relate to sagittal, axial and coronal sections. They are obtained by a calculation known per se, in particular by interpolation, from real sections acquired according to the imaging mode considered. In the example illustrated on the display screen 18 of [Fig. 1], a lower left window displays a three-dimensional representation, obtained by scanner, in side view of a bony part of a patient's ear. An upper left window displays a sagittal planar reconstruction by scanner including a positioning of the functional end 22 of the surgical instrument 20. An upper right window displays a coronal planar reconstruction by scanner including a positioning of the functional end 22 of the surgical instrument 20.Finally, a lower right window displays an axial plane reconstruction by scanner including a positioning of the functional end 22 of the surgical instrument 20. Of course, the aforementioned displays are interchangeable. In each window including a positioning of the functional end 22 of the surgical instrument 20, the latter is precisely located at the intersection of two axes, one horizontal and the other vertical. The fact of visualizing and following the functional end 22 simultaneously in the three sagittal, coronal and axial sections and in one or more imaging modes (for example scanner and / or MRI) is a valuable aid for the practitioner.

[0069] The main box 16 further comprises a software module 60, intended to be executed by the processing unit 44, for tracking by localization in real time of the functional end 22 of the surgical instrument 20 and for updating at each instant each of the planar reconstructions displayed simultaneously on the screen 18 which comprise it.

[0070] The main box 16 further comprises a software module 62, intended to be executed by the processing unit 44, for switching the display on the screen 18 between several planar sectional reconstructions of the object 12 according to a first three-dimensional representation (for example by scanner), several planar sectional reconstructions of the object 12 according to a second three-dimensional representation (for example by MRI) and several planar sectional reconstructions of the object 12 according to several three-dimensional representations (for example by scanner and MRI), this switching being in particular activatable by a predetermined proximity between the functional end 22 of the surgical instrument 20 and a specific receiver of the referencing ancillary 24, for example its calibration pad. This switching can also be activatable by direct interaction with the display screen 18 when it is touch-sensitive.Each window displayed on the screen 18 and presenting a planar reconstruction in sagittal, axial or coronal section of the object 12 can furthermore possibly be switched individually, for example by tactile interaction or other means, to change section and / or imaging mode.

[0071] The software modules 58, 60 and 62 are executable independently by the processing unit 44, not necessarily in the order of their presentation above. They can also be executed at the same time.

[0072] A method for three-dimensional resetting of at least one three-dimensional representation of the object 12 on the object 12 itself as it can be implemented by the navigation aid device of [Fig.l], corresponding to the execution of the software modules 50, 52, 54 and 56, will now be detailed in accordance with the succession of steps of [Fig.7].

[0073] During a step 100 carried out by executing the software module 50, a first three-dimensional representation of the object 12 according to a first three-dimensional imaging mode, for example according to a scanner acquisition mode, is recorded in the data storage area 48.

[0074] During a following step 102 carried out by execution of the software module 54 and / or by interaction, at least three predetermined reference points of the object 12 are located in the first three-dimensional representation and recorded in the data storage area 48. In the example of [Fig.7], which repeats that of [Fig.4], these are three reference points REF1, REF2, REF3 which are predetermined and located in a triangle around an area of ​​interest.

[0075] During a following step 104 carried out by executing the functional module 52-1 of the software module 52, an extraction of a surface of the object 12 in the first three-dimensional representation begins with a preliminary processing of this three-dimensional representation to distinguish at least one internal surface and one external surface of the object 12. As indicated previously, this can be done by thresholding, for example by binarization, of the three-dimensional representation on the basis of a predetermined gray level threshold value, which can be determined automatically or can be adjusted. This produces a “raw” three-dimensional representation such as that illustrated in the upper part of [Fig.6] (repeated in top view in [Fig.7]).

[0076] During subsequent steps 106 to 114 carried out by executing the functional module 52-2 of the software module 52, the surface extraction of the object 12 in the first three-dimensional representation continues by selecting at least one viewpoint external to the object 12 in its three-dimensional representation, preferably several viewpoints.

[0077] By way of non-limiting example, step 106 consists of determining a plane geometric figure representative of the reference points, for example the triangle T that they constitute in the case of the reference points REF1, REF2 and REF3 (cf. [Fig.4]).

[0078] Step 108 consists of determining an axis orthogonal to this plane geometric figure passing through a remarkable point, for example an isobarycenter, of this plane geometric figure. In the case of the triangle T defined by the points REF1, REF2 and REF3, it may be the orthogonal axis passing through its center of gravity.

[0079] Step 110 consists of determining a first viewpoint defined as belonging to this orthogonal axis, external to the object 12 and located at a first predefined distance from the remarkable point, for example the viewpoint PV1 at distance dl from RQ as illustrated in [Fig.4].

[0080] Step 112 consists of defining a reference surface from the first point of view PV1, for example the plane passing through PV 1 and parallel to the triangle T.

[0081] Finally, step 114 consists of automatically selecting other viewpoints defined as belonging to this reference surface and located at a second predefined distance from the first viewpoint in at least one direction, for example the viewpoints PV2, PV3, PV4 and PV5 at a distance d2 from PV 1 as illustrated in [Fig.4],

[0082] During a following step 116 carried out by executing the functional module 52-3 of the software module 52, the surface extraction of the object 12 in the first three-dimensional representation ends with the deletion of the internal surface(s) of the object 12 in the three-dimensional representation after thresholding 104 on a criterion of invisibility of this or these internal surface(s) from each selected point of view. For example, a result similar to that illustrated by [Fig.5] and the lower part of [Fig.6] is obtained in which only the external surface of the object 12 remains taken into account in the first three-dimensional representation (shown in top view in [Fig.7]). This external surface is recorded in the data storage area 48.

[0083] During a following step 118 carried out by execution of the software module 54 and by interaction, the reference points REF1, REF2 and REF3 previously identified in the first three-dimensional representation of the object 12 are located in the three-dimensional reference system linked to the referencing ancillary 24 using the tip 30 of the stylus 28 in a manner known per se and they are also recorded in the data storage area 48. In otology, this step can only take place after incision, local removal of the skin and fixing of the referencing ancillary 24 on the bone part by which an intervention is planned. On this occasion, the reference points REF1, REF2 and REF3 can again be located more finely in the three-dimensional representation of the external surface of the object 12 as obtained at the end of steps 106 to 116.The angular resetting of the first three-dimensional representation of the object 12 on the object 12 itself can then be calculated in a manner known per se.

[0084] During a following step 120 carried out by execution of the software module 56 and by interaction, several surface points Ps (represented in gray inside the circle represented by short broken lines), in particular other than the reference points REF1, REF2, REF3, are located on the surface of the object 12 in the three-dimensional reference system linked to the referencing ancillary 24 using the tip 30 of the stylus 28 and are recorded in the data storage area 48. Here again, the location of the surface points Ps is done directly in contact with the bone part of interest after local removal of the skin.

[0085] During a following step 122 carried out by executing the software module 56, the three-dimensional surface registration of the first three-dimensional representation is calculated from this recording of locations of surface points Ps. The calculation of the first passage matrix that it involves is facilitated by the prior angular registration carried out by executing step 118.

[0086] During a following step 124, the registration obtained from the first three-dimensional representation of the object 12 on the object 12 itself can be improved using the stylus 28 according to a counter-registration protocol known per se, so that this step will not be detailed.

[0087] During a following optional step 126 carried out by executing the software module 50 and equivalent to step 100, a second three-dimensional representation of the object 12 according to a second three-dimensional imaging mode, for example according to an MRI acquisition mode, is recorded in the data storage area 48.

[0088] During a following optional step 128 carried out by executing the software module 54 and equivalent to step 102, the predetermined reference points REF1, REF2, REF3 of the object 12 are located in the second three-dimensional representation and recorded in the data storage area 48. An angular resetting similar to that of step 118 can then be calculated by possibly using again the locations of the reference points REF1, REF2, REF3 determined using the tip 30 of the stylus 28 in step 118.

[0089] During a following optional step 130 carried out by executing the software module 52, similar to the succession of steps 104 to 116 previously detailed, an external surface of the object 12 is extracted in the second three-dimensional representation and is recorded in the data storage area 48. During this step, the respective positions of the different viewpoints PV1, PV2, PV3, PV4 and PV5 obtained at the end of step 114 can be recalculated by carrying out a recalibration between the reference points REF1, REF2, REF3 located at step 102 in the first three-dimensional representation and the same reference points REF1, REF2, REF3 located at step 128 in the second three-dimensional representation.

[0090] During a following optional step 132 carried out by executing the software module 56, the three-dimensional surface registration of the second three-dimensional representation is advantageously calculated by exploiting the recording of the locations of the surface points Ps carried out in step 120. The calculation of the second passage matrix that it involves is facilitated by the prior angular registration carried out by executing step 128.

[0091] Finally, during a following step 134, the registration obtained from the second three-dimensional representation of the object 12 on the object 12 itself can also be improved using the stylus 28 according to a counter-registration protocol identical to that of step 124.

[0092] It appears clearly that a method of three-dimensional registration of a three-dimensional representation of an object on the object itself such as that described previously makes it possible to obtain a completely satisfactory result even in difficult application contexts linked to a certain complexity of the surface of the object, in particular when the three-dimensional representation comprises at least one internal surface of the object, including any complex internal structure, in addition to its external surface.

[0093] It will also be noted that the invention is not limited to the embodiment described above.

[0094] In particular, the implementation of a three-dimensional registration method according to the invention has been described in a navigation aid device such as the surgical navigation device “Collin Navigation Solutions”. But according to another embodiment, the same three-dimensional registration method could be implemented in a robotic device for ENT or neurosurgical intervention such as the device marketed under the registered trademark “RobOtol” and also intended primarily for otorhinolaryngology interventions.

[0095] Indeed, such a robotic device may include a navigation aid device equipped with the elements essential for the implementation of the present invention, namely: - a memory for storing at least one three-dimensional representation of the object; - a computing unit designed for three-dimensional recalibration of this three-dimensional representation of the object on the object itself; - a referencing ancillary intended to be attached to the object without any degree of freedom; and - a pointer, which can be manipulated by a practitioner to carry out the recalibration, one tip of which can be located in a three-dimensional reference system linked to the referencing ancillary.

[0096] Furthermore, the robotic device has the advantage of comprising a robotic arm at the end of which a surgical instrument with a functional end can be attached, so that the movement of the latter can be controlled remotely by the practitioner.

[0097] It will be more generally apparent to those skilled in the art that various modifications may be made to the embodiment described above, in light of the teaching just disclosed to them. In the detailed presentation of the invention given above, the terms used should not be interpreted as limiting the invention to the embodiment set forth in the present description, but should be interpreted to include all equivalents the prediction of which is within the reach of those skilled in the art by applying their general knowledge to the implementation of the teaching just disclosed to them.

Claims

Claims

1. Method for three-dimensional recalibration of a three-dimensional representation of an object (12) on the object (12) itself, using a referencing ancillary (24) secured to the object (12) without a degree of freedom and a pointer (28) of which a tip (30) can be located in a three-dimensional reference frame linked to the referencing ancillary (24), comprising: - a recording (100, 126) in memory (48) of a three-dimensional representation of the object (12); - an extraction (104, 106, 108, 110, 112, 114, 116, 130), by a calculation unit (44), of a surface of the object (12) in the three-dimensional representation; and - a three-dimensional surface registration (120, 122, 132) of this three-dimensional representation on the object (12) comprising: • a recording (120), using the tip (30) of the pointer (28), of locations of a plurality of surface points (Ps) on the surface of the object (12) in the three-dimensional reference system linked to the referencing ancillary (24), and • a calculation (122, 132), by the calculation unit (44), of the three-dimensional surface registration from the extracted surface of the object (104, 106, 108, 110, 112, 114, 116, 130) in the three-dimensional representation and the recording (120) of locations of surface points; characterized in that the surface extraction (104, 106, 108, 110, 112, 114, 116, 130) comprises: - a preliminary processing (104, 130) of the three-dimensional representation to distinguish at least one internal surface and one external surface of the object (12); - a selection (106, 108, 110, 112, 114, 130) of at least one point of view (PV1, PV2, PV3, PV4, PV5) external to the object (12) in the three-dimensional representation; and - a deletion (116, 130) of said at least one surface internal in the three-dimensional representation on a criterion of invisibility of this at least one internal surface from each selected point of view (PV1, PV2, PV3, PV4, PV5).

2. A three-dimensional registration method according to claim 1, wherein: - the pre-processing (104, 130) comprises thresholding (104), for example binarization, of the three-dimensional representation on the basis of a predetermined gray level threshold value, automatically determinable or adjustable; and / or - the deletion (116, 130) of said at least one internal surface comprises obtaining a rendering of the object (12) in the pre-processed three-dimensional representation (104, 130) from each selected viewpoint (PV1, PV2, PV3, PV4, PV5), then a selection of the visible parts of this rendering of the object (12) from each selected viewpoint by a color buffer selection method.

3. Three-dimensional registration method according to claim 1 or 2, further comprising: - a recording (102, 128) in memory (48) of locations of at least three reference points (REF1, REF2, REF3) of the object (12) in the three-dimensional representation; - an angular three-dimensional registration (118, 128), by the calculation unit (44), of the three-dimensional representation on the object (12) from a recording (118), using the tip (30) of the stylus (28), of locations of the reference points (REF1, REF2, REF3) of the object (12) in the three-dimensional reference system linked to the referencing ancillary (24).

4. Three-dimensional registration method according to claim 3, in which the selection (106, 108, 110, 112, 114, 130) of said at least one viewpoint (PV1, PV2, PV3, PV4, PV5) external to the object (12) comprises: - the determination (108) of at least one axis carrying at least a point of view from the reference points (REF1, REF2, REF3) and / or at least one other remarkable point of the object (12); and - the automatic selection (110) of said at least one point of view (PV1) defined as belonging to this at least one carrier axis and external to the object (12).

5. Three-dimensional registration method according to claim 4, wherein the selection (106, 108, 110, 112, 114, 130) of said at least one viewpoint (PV1, PV2, PV3, PV4, PV5) external to the object (12) comprises: - the determination (106) of a plane geometric figure (T) representative of the reference points (REF1, REF2, REF3); - the determination (108) of an axis orthogonal to this plane geometric figure (T) passing through a remarkable point (RQ), for example an isobarycenter, of this plane geometric figure; - the automatic selection (110) of at least one first viewpoint (PV1) defined as belonging to this orthogonal axis, external to the object (12) and located at least a first predefined distance (dl) from the remarkable point (RQ); - the definition (112) of at least one reference surface from said at least one first point of view (PV1);and - the automatic selection (114) of other viewpoints (PV2, PV3, PV4, PV5) defined as belonging to this at least one reference surface and located at least a second predefined distance (d2) from said at least one first viewpoint (PV1) in at least one direction.;

6. Three-dimensional registration method according to claim 5, in which: - three reference points (REF1, REF2, REF3) are recorded (102, 128) and the plane geometric figure (T) representative of these three reference points is the triangle that they form; - the remarkable point (RQ) is one of the elements of the set consisting of the center of gravity of this triangle, the center of the circle inscribed in this triangle, the center of the circle circumscribed at this triangle and the orthocenter of this triangle; - a single first viewpoint (PV1) is defined, located at a single first predefined distance (dl) from the remarkable point (RQ); - this first predefined distance (dl) is either a constant or a variable proportional to the surface of this triangle or to at least one dimension of the object (12); - the reference surface is a plane or a three-dimensional conical surface, of the ellipsoid, paraboloid, hyperboloid type, passing through the first viewpoint; - the second predefined distance (d2) is either a constant or a variable proportional to the surface of this triangle or to at least one dimension of the object (12); and - two other viewpoints (PV2, PV4; PV3, PV5), respectively four other viewpoints (PV2, PV3, PV4, PV5), are selected (114) from the first viewpoint (PV1) in one direction, respectively two different directions.

7. Three-dimensional registration method according to any one of claims 3 to 6, in which the three-dimensional surface registration (120, 122, 132) of the three-dimensional representation on the object (12) comprises the recording (120) of the locations of several hundred surface points (Ps) on the surface of the object (12) in the three-dimensional reference system linked to the referencing ancillary (24), preferably between 200 and 1000 surface points, more preferably between 500 and 600 surface points.

8. Three-dimensional registration method according to any one of claims 1 to 7, in which the object (12) is a bone object, in particular a portion of temporal bone of a human or animal being.

9. Computer program downloadable from a communications network and / or recorded on a computer-readable medium and / or executable by a processor, comprising instructions (50, 52, 54, 56) for executing the steps of a method for three-dimensional registration of a three-dimensional representation of an object (12) on the object (12) itself according to any one of claims 1 to 8, when said program is executed by a calculation unit (44) of a device (10) for assisting navigation in the object (12) using a instrument (20) with functional end (22), and when this device (10) further comprises: - a memory (48) for storing a three-dimensional representation of the object (12); - a referencing ancillary (24) intended to be made integral with the object (12) without a degree of freedom; and - a pointer (28), manipulable by an operator, a tip (30) of which can be located in a three-dimensional reference system linked to the referencing ancillary (24).

10. Device (10) for assisting navigation in an object (12) using an instrument (20) with a functional end (22), comprising: - a memory (48) for storing a three-dimensional representation of the object (12); - a calculation unit (44) designed for three-dimensional registration of this three-dimensional representation of the object (12) on the object (12) itself; - a referencing ancillary (24) intended to be made integral with the object (12) without any degree of freedom; - a pointer (28) of which a tip (30) can be located in a three-dimensional reference system linked to the referencing ancillary (24); wherein the computing unit (44) is more precisely designed (50, 52, 54, 56) to execute: - an extraction (52) of a surface of the object (12) in the three-dimensional representation; and - a three-dimensional surface registration (56) of this three-dimensional representation on the object (12) comprising: • a recording, by interaction with the tip (30) of the pointer (28) when it is manipulated by an operator, of locations of a plurality of surface points (Ps) on the surface of the object (12) in the three-dimensional reference frame linked to the referencing ancillary (24), and • a calculation of the three-dimensional surface registration at from the extracted object surface (102) in the three-dimensional representation and from the recording of surface point locations; characterized in that the calculation unit (44) is further designed (52) so that the surface extraction comprises: - a preliminary processing (52-1) of the three-dimensional representation to distinguish at least one internal surface and one external surface of the object (12); - a selection (52-2) of at least one point of view (PV1, PV2, PV3, PV4, PV5) external to the object (12) in the three-dimensional representation; and - a deletion (52-3) of said at least one internal surface in the three-dimensional representation on a criterion of invisibility of this at least one internal surface from each selected point of view (PV1, PV2, PV3, PV4, PV5).

Images