CN110720882A - Endoscope system and endoscope operation control method - Google Patents
Endoscope system and endoscope operation control method Download PDFInfo
- Publication number
- CN110720882A CN110720882A CN201910984522.2A CN201910984522A CN110720882A CN 110720882 A CN110720882 A CN 110720882A CN 201910984522 A CN201910984522 A CN 201910984522A CN 110720882 A CN110720882 A CN 110720882A
- Authority
- CN
- China
- Prior art keywords
- endoscope
- angle
- rotation angle
- head end
- thumb wheel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
- A61B1/04—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor combined with photographic or television appliances
- A61B1/045—Control thereof
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
- A61B1/005—Flexible endoscopes
- A61B1/0051—Flexible endoscopes with controlled bending of insertion part
Abstract
The invention discloses an endoscope system, which can realize the automatic identification and control of the extending position and the extending angle of the head end of an endoscope, so that an operator does not need to analyze an endoscope image, namely the operator does not need to have rich endoscope image knowledge, thereby reducing the threshold and the difficulty of endoscope operation and further being beneficial to the popularization of an endoscopy technology; moreover, the accuracy that this application automatic identification and control head end stretched into the position and stretched into the angle is higher. The invention also discloses an endoscope operation control method, which has the same beneficial effects as the endoscope system.
Description
Technical Field
The present invention relates to the field of endoscopy, and in particular, to an endoscope system and an endoscope operation control method.
Background
The endoscope is a detection instrument integrating traditional optics, ergonomics, precision machinery, modern electronics, mathematics and software, and extends into the body through the natural orifice of the human body to perform detection. Because the natural cavities of the human body are very complicated and are bent in various ways, a doctor needs to continuously operate the manual thumb wheel in the process of endoscopy or treatment at present, so that the moving angle of the head end of the endoscope is adjusted, and the barrier-free movement of the endoscope is realized.
In the prior art, the angle adjustment process of the endoscope head end includes: acquiring an image of the current position of the head end of the endoscope for an operator to view; based on the experience of an operator, the dial wheel used for controlling the movement or rotation of the endoscope head end is manually adjusted according to the images so as to continuously adjust the movement angle of the endoscope head end. Therefore, the angle adjustment of the head end of the endoscope requires an operator to have rich endoscope image knowledge, so that the requirement on the operator is strict; moreover, the rotation angle of the thumb wheel determined in this way may be inaccurately adjusted, thereby affecting the precision of the movement angle of the endoscope head.
Therefore, how to provide a solution to the above technical problem is a problem that needs to be solved by those skilled in the art.
Disclosure of Invention
The invention aims to provide an endoscope system and an endoscope operation control method, which do not need an operator to analyze an endoscope image, thereby reducing the threshold and difficulty of endoscope operation and further being beneficial to the popularization of an endoscopy technology; moreover, the accuracy that this application automatic identification and control head end stretched into the position and stretched into the angle is higher.
In order to solve the technical problems, the invention provides an endoscope system, which comprises an endoscope body, a light source and a host, wherein the host is provided with a controller, the endoscope body comprises an insertion part which is inserted into a body cavity to be detected, one end of the insertion part is connected with the head end of an endoscope, the other end of the insertion part is connected with an operation end, the head end of the endoscope is provided with an image sensor, and the operation end is provided with a dial wheel; it still includes: the endoscope head end operation control system realizes automatic identification and control of the head end stretching position and the stretching angle.
Preferably, the endoscope head end operation control system acquires a current image of the endoscope head end through the image sensor, and identifies the head end target insertion position and the insertion angle through an angle planning model of the endoscope head end operation control system.
Preferably, the endoscope head end operation control system further comprises a display screen, the head end target extending position and extending angle identified by the endoscope head end operation control system are converted into a target rotation angle, and the display screen displays the target rotation angle, so that an operator can operate the dial wheel based on the target rotation angle.
Preferably, the endoscope head end operation control system further comprises a driving device for driving the dial wheel to rotate, and after the head end target extending position and the extending angle identified by the endoscope head end operation control system are converted into a target rotation angle, the driving device is controlled to automatically rotate to the target rotation angle.
Preferably, the driving device comprises a motor and a sensor which are connected with a thumb wheel, and the sensor is used for acquiring the actual rotation angle of the thumb wheel;
and the endoscope head end operation control system adopts the absolute value of the corner to control the motor to rotate to the target rotation angle or adopts the relative value of the corner to control the motor to rotate to the target rotation angle.
Preferably, the method steps of the endoscope head operating control system include:
an angle planning model for determining a standard rotation angle of a thumb wheel according to an endoscope image is established in advance and stored with an endoscope local machine;
and inputting the endoscope image into the angle planning model, identifying to obtain the head end target extending position and the extending angle, and converting to obtain the target rotation angle.
Preferably, after the target rotation angle is obtained, when a new target rotation angle of the thumb wheel is obtained in a current period, a difference value between the new target rotation angle and the target rotation angle of the thumb wheel obtained in a previous period is obtained, so as to obtain a rotation angle difference value, and thus the thumb wheel is operated to the new target rotation angle according to the rotation angle difference value.
Preferably, the training process of the angle planning model includes:
in the process of manually adjusting the head end of the endoscope, acquiring a sample image of the current position of the head end of the endoscope, and acquiring a standard rotation angle of the dial wheel;
inputting the sample image into the angle planning model to obtain a learning rotation angle of the thumb wheel;
and solving the error between the learning rotation angle and the standard rotation angle, and adjusting the parameters of the angle planning model according to the error so as to carry out the next round of sample learning until the error is reduced to the preset error range.
Preferably, the training process of the angle planning model includes:
extracting a sample image from a big data system in which an endoscope image is stored, and acquiring a standard rotation angle of the thumb wheel corresponding to the sample image;
and taking the sample image and the corresponding standard rotation angle thereof as input vectors, and performing unsupervised pre-training and supervised fine tuning on the angle planning model by using the input vectors to obtain a trained angle planning model.
In order to solve the above technical problem, the present invention further provides an endoscope operation control method, including:
acquiring an image of the current position of the head end of the endoscope;
and determining a target rotation angle of the thumb wheel corresponding to the image so as to operate the thumb wheel according to the target rotation angle.
The invention provides an endoscope system which can realize automatic identification and control of the extending position and the extending angle of the head end of an endoscope, so that an operator does not need to analyze an endoscope image, namely the operator does not need to have rich endoscope image knowledge, thereby reducing the threshold and difficulty of endoscope operation and further being beneficial to popularization of an endoscopy technology; moreover, the accuracy that this application automatic identification and control head end stretched into the position and stretched into the angle is higher.
The invention also provides an endoscope operation control method, which has the same beneficial effects as the endoscope system.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed in the prior art and the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a schematic view of an endoscopic system provided by an embodiment of the present invention;
fig. 2 is a flowchart of an endoscope head end operation control method according to an embodiment of the present invention;
FIG. 3 is a schematic view of a first endoscope head end operation control device according to an embodiment of the present invention;
FIG. 4 is a schematic view of a second endoscope head end operation control device according to an embodiment of the present invention;
FIG. 5 is a schematic view of a third endoscope head end operation control device provided in the embodiment of the present invention;
FIG. 6 is a first block diagram of an angle planning model according to an embodiment of the present invention;
fig. 7 is a second framework diagram of the angle planning model according to the embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Currently, in the process of endoscopy or treatment, a doctor needs to continuously operate the manual thumb wheel so as to adjust the moving angle of the head end of the endoscope, and therefore the unobstructed movement of the endoscope is realized. Because the natural cavities and ducts of the human body are complicated and various in curvature; if the head end of the endoscope does not extend into the natural cavity of the human body from the correct angle, the damage to the internal tissues of the human body and medical accidents are easily caused.
In the prior art, a doctor judges the position and the angle of a natural cavity according to an image of the current position of the endoscope head end displayed on a display and by combining the experience of the doctor, and manually adjusts a thumb wheel for controlling the movement or rotation of the endoscope head end, so that the endoscope head end can smoothly extend into the natural cavity, and therefore, the endoscopy or treatment in the prior art requires the doctor to have rich endoscope image knowledge; in order to reduce the damage to internal tissues caused by the fact that the head end of the endoscope extends into a human body, doctors are required to follow a strict operation flow; however, since there is a certain probability that the doctor operates manually, it is impossible to smoothly insert 100% of the endoscope tip into the human body and to cause no damage to the tissues.
Based on this, the inventor finds out the best natural orifice extending angle through automatic identification by modifying the prior art and through an artificial intelligence algorithm, and continuously optimizes learning through the algorithm by manually adjusting the thumb wheel or automatically controlling and adjusting the motor on the thumb wheel according to the best natural orifice extending angle, thereby replacing manual operation, avoiding the uncertainty of manual operation and reducing the damage to tissues.
The core of the invention is to provide an endoscope system and an endoscope operation control method, the optimal natural orifice extending angle is found through automatic identification, and an operator does not need to analyze an endoscope image, so that the threshold and the difficulty of endoscope operation are reduced, and the popularization of an endoscopy technology is facilitated; secondly, according to the endoscope head end moving angle automatic identification device, the dial wheel can be automatically operated and controlled, automatic control is achieved, and therefore the moving angle precision of the endoscope head end is improved.
The embodiments of the present invention will be further explained with reference to the drawings.
Referring to fig. 1, fig. 1 is a schematic view of an endoscope system according to an embodiment of the present invention.
The endoscope system comprises an endoscope body, a light source 100 and a host 200; wherein the endoscope body generally comprises an insertion part 300 for inserting into a body cavity to be detected, the insertion part 300 comprises a bendable soft endoscope body, and the endoscope body comprises an outer layer elastic tube and an inner layer snake bone; one end of the soft endoscope body of the insertion part 300 is connected with the endoscope head end 302, the other end of the soft endoscope body is connected with the operation end 301, two shifting wheels are arranged on the operation end 301, the shifting wheels penetrate through the inner layer snake bone through a traction steel wire rope to be connected onto the endoscope head end 302, and therefore the two shifting wheels respectively control the moving angles of the endoscope head end 302 in the up-down direction and the left-right direction.
The light source 100 may include a light source for generating illumination light, and may also include a light source for generating special light for special illumination; the light source can be in the form of a single light source, a plurality of light sources or a multi-path light source combined beam; the light source type includes, but is not limited to, a xenon light source, an LED light source, or a laser light source.
An aqueous vapor system and an imaging system are further disposed on the endoscope head base of the endoscope head 302, the imaging system is configured with an imaging element of the observation optical system, and the imaging element may be a CCD or a CMOS (Complementary metal oxide Semiconductor) or other elements; the camera element sends image signals to the host 200 for image data processing, and feeds back part or all of the data results to the light source 100 and/or the display; the display may include, but is not limited to: LCD displays, LED displays, OLED displays, quantum dot displays, laser displays, and the like.
The host 200 has a driving circuit, a controller and/or an image processor, the driving circuit can drive the lighting circuit and/or drive the electric thumb wheel; the image processor can perform image processing on the image signal fed back by the camera element, and the controller is connected with the output end of the image processor and used for identifying the extending position and the extending angle of the head end of the endoscope after receiving the image and converting the extending position and the extending angle of the head end of the endoscope into a target rotating angle for determining the dial wheel at the moment; the image processor and the controller can be two independent processors, or can realize two functions of image processing and control for one processor.
Referring to fig. 2, fig. 2 is a system block diagram of an endoscope head end operation control method according to an embodiment of the present invention.
The process of determining the target rotation angle of the jog dial 1 corresponding to the image includes:
an angle planning model for determining a standard rotation angle of the thumb wheel 1 according to an endoscope image is established and trained in advance;
inputting an endoscope image into an angle planning model, and identifying the extending position and the extending angle of the head end of the endoscope; and the processor of the endoscope converts the extending position and the extending angle of the head end of the endoscope into a target rotating angle for determining the thumb wheel 1.
Specifically, in order to determine the target rotation angle of the thumb wheel 1 corresponding to the endoscope image, an angle planning model for determining the standard rotation angle of the thumb wheel 1 according to the endoscope image is established in advance (namely the input of the angle planning model is the endoscope image, the output is the endoscope head end extending position and the extending angle corresponding to the endoscope image, a processor of the endoscope converts the endoscope head end extending position and the extending angle into the target rotation angle for determining the thumb wheel 1), and then the angle planning model is trained to obtain a more accurate angle planning model, so that the angle planning model is used for establishing the corresponding relation between the endoscope image and the target rotation angle of the thumb wheel 1.
As an alternative embodiment.
Referring to fig. 3, fig. 3 is a schematic diagram of a first endoscope head end operation control device according to an embodiment of the present invention.
The endoscopic operation control device may include:
and a display 4 connected to the controller 3.
Further, the endoscope operation control device of the present application may further include a display screen 4 controlled by the controller 3, and the operation principle thereof is:
during the process of endoscopy or treatment, a doctor inserts the insertion part of the endoscope into a natural cavity of a human body, and during the process of moving the head end of the endoscope, the image sensor 2 arranged at the head end of the endoscope can acquire an image (an endoscope image for short) of the current position of the head end of the endoscope; a processor of the endoscope identifies the extending position and the extending angle of the head end of the endoscope through the endoscope image; and the processor of the endoscope converts the extending position and the extending angle of the head end of the endoscope into the target rotating angle of the determined thumb wheel 1 at the moment.
The endoscope automatically identifies the extension position and the extension angle of the head end of the endoscope, converts the extension position and the extension angle of the head end of the endoscope into a target rotation angle for determining the dial wheel 1 at the moment, can display related information on the display screen 4, plays a role in guiding an operator to operate the dial wheel 1, and the operator can operate the dial wheel 1 based on the displayed target rotation angle, so that the threshold and the difficulty of endoscope operation are reduced.
Except that the display shows, the endoscope operation control device of this application can also add other devices (like the voice broadcast ware, the voice broadcast ware is connected with controller 3, controller 3 still is used for controlling the voice broadcast ware to report target turned angle, supplies the operator to listen) that are used for instructing operator operation thumb wheel 1.
As an alternative embodiment.
Referring to fig. 4, fig. 4 is a schematic diagram of a second endoscope head end operation control device according to an embodiment of the present invention.
The endoscope operation control device further includes:
the driving device 5 is connected with the controller 3 and is used for driving the thumb wheel 1 to rotate;
the controller 3 is also used for automatically controlling the driving device 5 to drive the thumb wheel 1 to rotate according to the target rotation angle.
Further, the endoscope operation control device of the present application may further include an operation device 5 controlled by the controller 3, and the operation principle thereof is:
considering that the moving environment of the head end of the endoscope is complex, the action of the operator for operating the thumb wheel 1 is frequent, and the operation workload of the operator is large, the driving device 5 which can be controlled by the controller 3 and is used for driving the thumb wheel 1 to rotate is additionally arranged in the endoscope, so that a foundation is laid for controlling the automatic rotation of the thumb wheel 1.
More specifically, since the controller 3 of the present embodiment controls the driving device 5 to target that the driving device 5 drives the thumb wheel 1 to rotate at the target rotation angle, after the controller 3 obtains the target rotation angle of the thumb wheel 1 corresponding to the endoscope image, the driving device 5 is automatically controlled to drive the thumb wheel 1 to rotate based on the control target of the driving device 5, so that the driving device 5 drives the thumb wheel 1 to automatically rotate at the target rotation angle, and the operation workload of the operator is reduced.
The driving means 5 includes, but is not limited to, a combination of a motor and a transmission means (e.g., a gear), a combination of a motor and a code wheel sensor, a step motor, a servo electrode, etc.
The driving device 5 may be installed inside the handle to protect the motor from external influences, or may be installed at other positions of the endoscope to save installation space.
The control method of the controller 3 for controlling the driving device 5 includes, but is not limited to, absolute value control and relative value control, the absolute value control is to control the motor to rotate to an absolute angle value, the relative value control is to calculate the difference between the actual rotation angle and the target rotation angle, and the motor is controlled to rotate to the target rotation angle by an incremental value of the control motor.
As an alternative embodiment.
Referring to fig. 5, fig. 5 is a schematic diagram of a third endoscope head end operation control device according to an embodiment of the present invention.
The endoscope operation control device may not include: a display screen 4.
Further, the endoscope operation control device of the present application does not include the display screen 4, and the working principle thereof is as follows:
during the process of endoscopy or treatment, a doctor inserts the insertion part of the endoscope into a natural cavity of a human body, and during the process of moving the head end of the endoscope, the image sensor 2 arranged at the head end of the endoscope can acquire an image (an endoscope image for short) of the current position of the head end of the endoscope; a processor of the endoscope identifies the extending position and the extending angle of the head end of the endoscope through the endoscope image; and the processor of the endoscope converts the extending position and the extending angle of the head end of the endoscope into the target rotating angle of the determined thumb wheel 1 at the moment.
The controller 3 is further configured to automatically control the driving device 5 to drive the dial wheel 1 to rotate according to the target rotation angle, and the control manner for controlling the driving device 5 to drive the dial wheel 1 to rotate is as described in the above embodiment.
As an alternative embodiment.
Referring to fig. 6, fig. 6 is a first frame diagram of an angle planning model according to an embodiment of the present invention.
The main body 200 of the endoscopic manipulation control apparatus further includes an internal memory, the angle planning model is stored in the internal memory by being pre-stored,
the process of determining the target rotation angle of the jog dial 1 corresponding to the image includes:
an angle planning model for determining a standard rotation angle of the thumb wheel 1 according to an endoscope image is established, stored and trained in advance;
inputting the image into an angle planning model to obtain the extending position and the extending angle of the head end of the endoscope; converting the extending position and the extending angle of the head end of the endoscope into a target rotating angle of the determined thumb wheel 1 by a processor of the endoscope;
and solving the difference between the new target rotation angle and the target rotation angle of the thumb wheel 1 obtained in the previous period to obtain a rotation angle difference.
Specifically, considering that the absolute rotation angle of the dial 1 (i.e. the angle to which the dial 1 should rotate from the initial rotation angle) is output by the angle planning model, in the process of actually operating the dial 1, the absolute rotation angle of the dial 1 output by the angle planning model needs to be different from the current rotation angle of the dial 1, and the rotation angle difference between the absolute rotation angle and the current rotation angle is the current angle to which the dial 1 should actually rotate (i.e. the rotation angle difference is a direct reference value for the rotation of the dial 1), especially for the dial 1 operated by the operator, if the absolute rotation angle of the dial 1 is taken as the reference value for the operator, the operator needs to calculate the current angle to which the dial 1 should actually rotate based on the current rotation angle of the dial 1. Based on this, make the difference through the turned angle, for operator's operation thumb wheel 1, the operator directly according to turned angle difference operation thumb wheel 1 can, need not to calculate to be convenient for the operator to operate.
More specifically, the target rotation angle of the thumb wheel 1 obtained in the previous period may be subtracted from the new target rotation angle in the current period (or the new target rotation angle of the thumb wheel 1 obtained in the previous period may be subtracted from the target rotation angle in the previous period, which is similar in principle and is not described herein again), and if the rotation angle difference is a positive value, the thumb wheel 1 is correspondingly rotated in the forward direction by a certain angle; if the difference value of the rotation angles is a negative value, the dial wheel 1 correspondingly rotates reversely for a certain angle.
The training process of the angle planning model comprises the following steps:
in the process of manually adjusting the head end of the endoscope, a sample image of the current position of the head end of the endoscope is obtained, and meanwhile, a standard rotation angle of the thumb wheel 1 is obtained;
inputting the sample image into an angle planning model to obtain a learning rotation angle of the thumb wheel 1;
and solving the error between the learning rotation angle and the standard rotation angle, adjusting the parameters of the angle planning model according to the error to perform the next round of sample learning until the error is reduced to a preset error range, and updating the final parameters of the angle planning model to a local memory.
Specifically, the sample training process of the angle planning model of the present application includes: firstly, a trained sample image is obtained (the sample image is an endoscope image obtained in the process that an operator manually operates the dial wheel 1 to move the endoscope head end), and meanwhile, a standard rotation angle of the dial wheel 1 corresponding to the sample image is obtained (namely, in the process that the operator manually operates the dial wheel 1, the rotation angle of the dial wheel 1 manually operated by the operator corresponding to the obtained sample image (obtained by using a code wheel sensor) is obtained, it can be understood that the standard rotation angle can be used as the standard output of an angle planning model in sample training), and then, the sample image is input to an angle planning model, so that the learning rotation angle of the dial wheel 1 is obtained by using the angle planning model. The training target of the angle planning model is that for the same sample image, the corresponding learning rotation angle is consistent with the standard rotation angle, so that the error between the learning rotation angle and the standard rotation angle is obtained, the parameters of the angle planning model are adjusted according to the error, then the next round of sample learning is carried out (namely, new sample images are obtained for continuous learning), and the sample training of the angle planning model is finished and the result is updated to the local memory until the error is reduced to be within the preset error range.
Further, considering that the head end of the endoscope can work in different mobile environments, the trained angle planning model may only be applicable to some relatively similar mobile environments, but cannot be applicable to mobile environments with large differences from the relatively similar mobile environments. Therefore, the method and the device can also check the trained angle planning model to verify the universality of the angle planning model.
Specifically, the verification process of the angle planning model includes: firstly, a verified new sample image is obtained (the new sample image is an endoscope image obtained in the process that an operator manually adjusts the dial wheel 1 to move the endoscope head end, it can be understood that the new sample image has a certain difference from the trained sample image), and a new standard rotation angle of the dial wheel 1 corresponding to the new sample image is obtained at the same time (that is, in the process that the operator manually adjusts the dial wheel 1, the obtained rotation angle of the dial wheel 1 corresponding to the new sample image is manually adjusted by the operator, it can be understood that the new standard rotation angle can be used as a standard output of the angle planning model in sample verification), and then the new sample image is input to the angle planning model, so that a new learned rotation angle of the dial wheel 1 is obtained by using the angle planning model. The training standard-reaching target of the angle planning model is that for the same new sample image, the corresponding new learning rotation angle and the new standard rotation angle are kept consistent (the error of the two is considered to be consistent in a preset error range), so that the error of the new learning rotation angle and the new standard rotation angle is obtained, whether the error is in the preset error range is judged, if the error is in the preset error range, the trained angle planning model meets the universality requirement, namely the verification is passed, the training of the angle planning model is really finished, and the angle planning model is put into use; if the error is not within the preset error range, the trained angle planning model is not in accordance with the universality requirement, namely the verification fails, the angle planning model is retrained by using the new sample image until the error is reduced to the preset error range, the retraining of the angle planning model is finished, then a new round of verification is carried out, and finally the angle planning model is in accordance with the universality requirement.
As an alternative embodiment.
Referring to fig. 7, fig. 7 is a second framework diagram of an angle planning model according to an embodiment of the present invention.
The main body 200 of the endoscope operation control device includes a controller, which can also realize automatic control operations including automatic control of the power jog dial, endoscope main body networking, remote update/control, and the like.
The endoscope local machine can be connected with a server of a big data platform through a host networking function of the controller, and is provided with a memory for storing endoscope images to be sent; the endoscope head end is recognized to extend into the position and the angle through the big data platform, the endoscope head end is converted into the target rotation angle for determining the dial wheel 1 at the moment according to the position and the angle, and the big data platform can acquire the endoscope images from different endoscope machines, so that the big data platform is high in training efficiency, wide in trainable range, capable of improving the accuracy of the operation, and rich in resources of the big data platform, and capable of performing various complex artificial intelligence algorithms.
The training process of the angle planning model comprises the following steps:
extracting a sample image from a big data system in which an endoscope image is stored, and acquiring a standard rotation angle of a thumb wheel 1 corresponding to the sample image;
and taking the sample image and the corresponding standard rotation angle thereof as input vectors, and performing unsupervised pre-training and supervised fine tuning on the angle planning model by using the input vectors to obtain the trained angle planning model.
Specifically, in addition to the above-mentioned training method for the angle planning model, the present application may also train the angle planning model using big data, and the training principle of training the angle planning model using big data is as follows:
in the process of manually operating the thumb wheel 1 by an operator, the acquired endoscope image and the rotation angle (as the standard rotation angle of the thumb wheel 1) of the thumb wheel 1 manually operated by the operator corresponding to the endoscope image are both stored in the big data system. When an angle planning model is trained, firstly, an endoscope image (as a sample image) is extracted from a big data system, and a standard rotation angle of a thumb wheel 1 corresponding to the sample image is obtained; and then, taking the sample image and the corresponding standard rotation angle thereof as input vectors, and carrying out unsupervised pre-training and supervised fine tuning on the angle planning model by using the input vectors to obtain the trained angle planning model. More specifically, the angle planning model comprises an encoding layer and a logistic regression layer, the sample image and the corresponding standard rotation angle thereof are input vectors of the angle planning model, and the rotation angle pre-measurement of the thumb wheel 1 is an output vector of the angle planning model; in an encoding layer, carrying out unsupervised pre-training by using an input vector to obtain high-level representation of a sample image; in a logistic regression layer, the rotation angle prediction quantity of the thumb wheel 1 is obtained by utilizing the high-level representation of the sample image, a loss function is established by utilizing the input vector, the rotation angle prediction quantity of the thumb wheel 1 and the parameters of the angle planning model, and the coding layer is used for adjusting the parameters of the angle planning model to minimize the loss function and carrying out supervised fine adjustment to obtain the trained angle planning model.
As an alternative embodiment.
The main unit 200 of the endoscope operation control device further comprises an internal memory, the angle planning model can be connected with the endoscope local machine and the server of the big data platform through the pre-storage and the internal memory and the host machine networking function of the controller, relevant parameters of the angle planning model can be obtained through the server training of the big data platform, and the parameters of the endoscope local machine are updated regularly.
The training efficiency based on the big data platform is high, the trainable range is wide, the accuracy of the operation is improved, the accuracy of local detection is guaranteed by updating local parameters, the mode not only keeps the high quality of local identification and detection, but also combines the advantages of the big data platform.
Finally, the dial wheel driving motor is automatically controlled through a controller of the endoscope, and the head end of the endoscope is driven to reach the extending angle, so that automatic control is realized; thereby leading the head end of the endoscope to move at a proper angle without obstruction. Therefore, the endoscope image analysis method and the endoscope image analysis system do not need an operator to analyze the endoscope image, namely the operator does not need to have rich endoscope image knowledge, so that the threshold and the difficulty of endoscope operation are reduced, and the popularization of the endoscope inspection technology is facilitated; in addition, the automatically determined target rotation angle operation dial wheel has high precision, so that the precision of the moving angle of the head end of the endoscope is improved.
It is further noted that, in the present specification, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. An endoscope system comprises an endoscope body, a light source and a host, wherein the host is provided with a controller, the endoscope body comprises an insertion part which is inserted into a body cavity to be detected, one end of the insertion part is connected with an endoscope head end, the other end of the insertion part is connected with an operation end, the endoscope head end is provided with an image sensor, and the operation end is provided with a dial wheel; it is characterized in that it also comprises: the endoscope head end operation control system realizes automatic identification and control of the head end stretching position and the stretching angle.
2. An endoscope system according to claim 1 and wherein said endoscope head end operation control system obtains current images of the endoscope head end through said image sensor, and identifies the head end target insertion position and insertion angle through its angle planning model.
3. An endoscope system according to claim 2 and also comprising a display screen, wherein said endoscope head end operation control system converts a target rotation angle into a head end target insertion position and insertion angle identified by said endoscope head end operation control system, and said display screen displays said target rotation angle to facilitate an operator to operate said thumb wheel based on said target rotation angle.
4. An endoscope system according to claim 2 or claim 3, further comprising a driving device for driving the thumb wheel to rotate, wherein the endoscope head end operation control system converts the head end target extending position and extending angle identified by the head end target extending position and extending angle into a target rotating angle, and controls the driving device to automatically rotate to the target rotating angle.
5. The endoscopic system of claim 4 wherein said drive means comprises a motor connected to a thumb wheel and a sensor for acquiring an actual angle of rotation of said thumb wheel;
and the endoscope head end operation control system adopts the absolute value of the corner to control the motor to rotate to the target rotation angle or adopts the relative value of the corner to control the motor to rotate to the target rotation angle.
6. An endoscope system according to claim 1 and wherein said endoscope head end method steps of operating a control system comprise:
an angle planning model for determining a standard rotation angle of a thumb wheel according to an endoscope image is established in advance and stored with an endoscope local machine;
and inputting the endoscope image into the angle planning model, identifying to obtain the head end target extending position and the extending angle, and converting to obtain the target rotation angle.
7. The endoscope system of claim 6 wherein after obtaining the target rotation angle, when obtaining a new target rotation angle of the thumb wheel in a current cycle, obtaining a difference between the new target rotation angle and the target rotation angle of the thumb wheel obtained in a previous cycle to obtain a rotation angle difference to achieve operation of the thumb wheel to the new target rotation angle according to the rotation angle difference.
8. An endoscope system according to any of claims 6-7 and wherein said angular programming model training procedure comprises:
in the process of manually adjusting the head end of the endoscope, acquiring a sample image of the current position of the head end of the endoscope, and acquiring a standard rotation angle of the dial wheel;
inputting the sample image into the angle planning model to obtain a learning rotation angle of the thumb wheel;
and solving the error between the learning rotation angle and the standard rotation angle, and adjusting the parameters of the angle planning model according to the error so as to carry out the next round of sample learning until the error is reduced to the preset error range.
9. An endoscope system according to any of claims 6-7 and wherein said angular programming model training procedure comprises:
extracting a sample image from a big data system in which an endoscope image is stored, and acquiring a standard rotation angle of the thumb wheel corresponding to the sample image;
and taking the sample image and the corresponding standard rotation angle thereof as input vectors, and performing unsupervised pre-training and supervised fine tuning on the angle planning model by using the input vectors to obtain a trained angle planning model.
10. An endoscope operation control method characterized by comprising:
acquiring an image of the current position of the head end of the endoscope;
and determining a target rotation angle of the thumb wheel corresponding to the image so as to operate the thumb wheel according to the target rotation angle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910984522.2A CN110720882A (en) | 2019-10-16 | 2019-10-16 | Endoscope system and endoscope operation control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910984522.2A CN110720882A (en) | 2019-10-16 | 2019-10-16 | Endoscope system and endoscope operation control method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110720882A true CN110720882A (en) | 2020-01-24 |
Family
ID=69221413
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910984522.2A Pending CN110720882A (en) | 2019-10-16 | 2019-10-16 | Endoscope system and endoscope operation control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110720882A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021180043A1 (en) * | 2020-03-09 | 2021-09-16 | 陈阶晓 | Endoscope image correction system and method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101455554A (en) * | 2007-12-10 | 2009-06-17 | 奥林巴斯医疗株式会社 | Endoscope system |
| CN105902253A (en) * | 2016-04-28 | 2016-08-31 | 深圳市鹏瑞智能图像有限公司 | Endoscope insertion control method and system |
| CN107361728A (en) * | 2017-08-16 | 2017-11-21 | 珠海康弘发展有限公司 | Endoscope guidance method and guider |
| CN108685560A (en) * | 2017-04-12 | 2018-10-23 | 香港生物医学工程有限公司 | Automation steering and method for robotic endoscope |
| TW201934073A (en) * | 2018-02-01 | 2019-09-01 | 台灣骨王生技股份有限公司 | Surgical instrument guiding system including a smart glasses |
-
2019
- 2019-10-16 CN CN201910984522.2A patent/CN110720882A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101455554A (en) * | 2007-12-10 | 2009-06-17 | 奥林巴斯医疗株式会社 | Endoscope system |
| CN105902253A (en) * | 2016-04-28 | 2016-08-31 | 深圳市鹏瑞智能图像有限公司 | Endoscope insertion control method and system |
| CN108685560A (en) * | 2017-04-12 | 2018-10-23 | 香港生物医学工程有限公司 | Automation steering and method for robotic endoscope |
| CN107361728A (en) * | 2017-08-16 | 2017-11-21 | 珠海康弘发展有限公司 | Endoscope guidance method and guider |
| TW201934073A (en) * | 2018-02-01 | 2019-09-01 | 台灣骨王生技股份有限公司 | Surgical instrument guiding system including a smart glasses |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021180043A1 (en) * | 2020-03-09 | 2021-09-16 | 陈阶晓 | Endoscope image correction system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR101977181B1 (en) | Dental intraoral scanner system | |
| EP2583616B1 (en) | Endoscope | |
| EP2215960B1 (en) | Endoscope curve control apparatus | |
| JP5301867B2 (en) | Medical manipulator system | |
| CN102469914B (en) | Endoscope apparatus and bending drive control method | |
| JP5559996B2 (en) | Endoscope device, endoscope system, and operation method of endoscope device | |
| EP2301411A1 (en) | Endoscope system | |
| JP2008194302A (en) | Endoscope system controlling its treatment tool position and its control method | |
| EP3395292B1 (en) | Three-dimensional-measurement method, and three-dimensional-measurement device | |
| US10542198B2 (en) | Operation support system, wearable apparatus, image pickup apparatus, and operation support method | |
| US20220175218A1 (en) | Information processing device, endoscope control device, information processing method and operating method of endoscope control device | |
| KR101983342B1 (en) | Apparatus for automatic intubation of artificial intelligence and method for driving the same | |
| EP3061418A1 (en) | Medical manipulator and method of initializing medical manipulator | |
| WO2020165978A1 (en) | Image recording device, image recording method, and image recording program | |
| CN110720882A (en) | Endoscope system and endoscope operation control method | |
| US20220218180A1 (en) | Endoscope insertion control device, endoscope insertion control method, and non-transitory recording medium in which endoscope insertion control program is recorded | |
| CN116543889B (en) | Endoscope remote control system and method based on medical Internet of things | |
| WO2019003272A1 (en) | Flexible tube insertion device, insertion control device, and insertion control program | |
| CN110549328A (en) | Job support control device and method, job image control device and display method | |
| JP2012020028A (en) | Processor for electronic endoscope | |
| CN107361728A (en) | Endoscope guidance method and guider | |
| JP7454112B2 (en) | Continuum robotic devices, methods and media | |
| CN112584735A (en) | Image correction of surgical endoscopic video streams | |
| EP4159161A1 (en) | Method and apparatus for predicting missing tooth in oral image | |
| KR102495838B1 (en) | Method, apparatus and computer program for controlling endoscope based on medical image |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200124 |
|
| RJ01 | Rejection of invention patent application after publication |