Systems, devices, and methods for turbidity analysis

Abstract

An endoscopy system includes an endoscopy imager configured to capture images of a target site in a living body; and a processor configured to determine one or more image metrics for each of a plurality of image frames captured over a time span, analyze changes in the image metrics over the time span, and determine a turbidity metric for the target site based on the analyzed changes in the image metrics.

Description

[0001] CLAIM OF PRIORITY

[0002] This disclosure claims priority to U.S. Provisional Patent Application Serial No. 62 / 904,882, filed September 24, 2019; the disclosure of which is incorporated herein by reference. TECHNICAL FIELD

[0003] The present disclosure relates to a system, device, and method for performing endoscopic surgery, and in particular, turbidity analysis of an endoscopic imaging environment. BACKGROUND

[0004] Endoscopic imagers can be used during a variety of medical interventions. When the imaging environment is hazy or cloudy, the anatomical view provided by the imager is limited. Cloudiness can be caused by blood, urine, or other particulates. In some endoscopic surgeries, such as ureteroscopy, the cloudy imaging environment can be managed by a fluid management system that circulates fluid in the imaged cavity. SUMMARY

[0005] The present disclosure relates to an endoscopic system comprising an endoscopic imager configured to capture image frames of a target site in a living body; and a processor. The processor is configured to: determine one or more image metrics for each of a plurality of image frames captured over a time span; analyze changes in the image metrics over the time span; and determine a turbidity metric of the target site based on the changes in the analyzed image metrics.

[0006] In one embodiment, the image metrics are image entropy metrics comprising a red entropy metric and a cyan entropy metric.

[0007] In one embodiment, the processor is further configured to: estimate a blood content in the current image frame; and alter the current image frame to mitigate a visual effect of the blood content and enhance a remainder of the current image frame.

[0008] In one embodiment, the processor is further configured to: identify particulates in the current image frame and classify the particulates in the current image frame.

[0009] In one embodiment, the endoscopic system further comprises a display configured to annotate the current image frame with the turbidity metric.

[0010] In one embodiment, the display is further configured to: group the identified particulates into a class; and annotate the current image frame using the particulate classification.

[0011] In one embodiment, the identified particulates are kidney stones, and the classification relates to a size of the kidney stones.

[0012] The present disclosure also relates to an endoscopy system comprising an endoscopic imager configured to capture image frames of a target site within a living body; a fluid delivery mechanism to provide irrigation fluid to the target site to clear a field of view of the endoscopic imager; a processor configured to: determine a turbidity metric for at least one image from the imager; determine a fluid delivery adjustment for the irrigation fluid based on the turbidity metric; and control the fluid delivery mechanism to adjust fluid delivery provided by the fluid delivery mechanism based on the determined fluid delivery adjustment.

[0013] In one embodiment, the processor is further configured to determine a type of the interventional activity based on the feature detection identifying the interventional instrument.

[0014] In one embodiment, the processor is further configured to determine a phase of the interventional activity; and adjust the fluid delivery provided by the fluid delivery mechanism based on the phase of the interventional activity.

[0015] In one embodiment, the processor is further configured to identify and classify particles in the current image frame; and adjust the fluid delivery provided by the fluid delivery mechanism based on the particle classification.

[0016] In one embodiment, the particles are kidney stones, and the particle classification relates to a size of the kidney stones.

[0017] In one embodiment, the processor is further configured to determine a blood metric for the at least one image; and adjust the fluid delivery provided by the fluid delivery mechanism based on the blood metric.

[0018] In one embodiment, the processor is further configured to determine an image entropy metric for the at least one image.

[0019] In one embodiment, the turbidity metric and the blood metric are determined based in part on the image entropy metric.

[0020] Further, the present disclosure relates to a method comprising determining one or more image metrics for each of a plurality of image frames of a target site within a living body captured over a time span; analyzing changes in the image metrics over the time span; and determining a turbidity metric for the target site based on the analyzed changes in the image metrics.

[0021] In one embodiment, the image metrics are image entropy metrics including a red entropy metric and a cyan entropy metric.

[0022] In one embodiment, the method further comprises estimating a blood content in the current image frame; and altering the current image frame to mitigate a visual effect of the blood content and enhance a remainder of the current image frame.

[0023] In one embodiment, the method further includes identifying particles in the current image frame and classifying the particles in the current image frame.

[0024] In one embodiment, the method further includes annotating the current image frame with a turbidity metric. Attached Figure Description

[0025] Figure 1 Systems for performing endoscopic surgery according to various exemplary embodiments of the present disclosure are shown.

[0026] Figure 2 Methods for managing irrigation fluid flow in a closed-loop feedback system according to various exemplary embodiments of the present invention are shown.

[0027] Figure 3 A method for enhancing visibility through a liquid medium turbid with blood, according to a first exemplary embodiment, is shown.

[0028] Figure 4 A method for enhancing visibility through a liquid medium turbid with blood, according to a second exemplary embodiment, is shown.

[0029] Figure 5 Description is shown Figures 2-4 A flowchart of a combination of methods. Detailed Implementation

[0030] This disclosure can be further understood with reference to the following description and the accompanying drawings, wherein similar elements are referred to by the same reference numerals. Exemplary embodiments describe algorithmic improvements for managing turbidity in an endoscopic imaging environment. Typical urological procedures utilize imaging devices (e.g., ureteroscopes or other endoscopic imagers), mechanisms for providing fluid (for clearing the field of view and / or expanding body cavities), and treatment mechanisms (e.g., Boston Scientific LithoVue). TM (Equipment and / or energy sources for lasers or radio frequencies, etc.).

[0031] The improvements described in this paper include, for example, methods for analyzing endoscopic images and determining turbidity, as well as providing information to fluid management systems to manage clarity within the field of view. Other improvements include algorithmically subtracting blood features from endoscopic images and enhancing image features blurred by blood, particularly for urological procedures. Some common urological procedures include treatment of kidney stones (e.g., lithotripsy), BPH (i.e., benign prostatic hyperplasia) surgery (e.g., GreenLight Cure). TMprostatectomy, bladder tumor resection, uterine fibroid treatment, diagnosis, etc., although those skilled in the art will appreciate that the equipment and techniques for improving the image can be used in a variety of procedures in which turbidity is an issue (i.e., non-urological procedures as well).

[0032] Figure 1 A system 100 for performing endoscopic surgery is shown in accordance with various exemplary embodiments of the present disclosure. The system 100 includes an endoscope 102 having an imager 104 for acquiring image frames of an anatomical site within a living body during an endoscopic procedure and a fluid delivery mechanism 106 for providing fluid (e.g., saline) to the anatomical site to clear blood and debris that can impair the field of view of the imager 104. The fluid delivery mechanism 106 also provides suction to simultaneously remove fluid from the anatomical site. In this way, the anatomical site is continually refreshed with a substantially transparent fluid so that clearer images can be generated.

[0033] The system 100 can also include a treatment device 108, which is selected according to the nature of the endoscopic procedure. The treatment device 108 can be run through the endoscope 102 or can be located externally to the endoscope 102. For example, the treatment device 108 can be, for example, a laser or shock wave generator for breaking up kidney stones or an electrotome for removing prostate tissue. When the endoscopic procedure is for diagnostic purposes, i.e., for examining the anatomical site rather than for treating a disease, the treatment device can not be used. Although exemplary embodiments are described with respect to urological imaging, exemplary embodiments are not limited thereto. Certain embodiments can also be applicable to a wide range of procedures, such as, for example, endoscopic procedures within the digestive system, etc., including endoscopic procedures that do not include a fluid delivery mechanism.

[0034] The system 100 includes a computer 110 that processes the image frames provided by the imager 104 and provides the processed images to a display 112. In the present embodiment, the computer 110 and the display 112 are provided at an integrated station, such as an endoscope console. Other features for performing urological surgical procedures can be implemented on the endoscope console, including, for example, actuators that control the flow rate, pressure, or dispensing of the fluid delivered through the fluid delivery mechanism 106. Exemplary embodiments describe algorithmic processes for altering and enhancing the displayed images, typically continuously or in any other desired manner.

[0035] In one embodiment, image metrics determined from the captured image frames are used to estimate the degree to which the video sequence is obscured by a turbid imaging environment. The image metrics include an image entropy metric. Image entropy can generally be defined as a measure of information content in an image, which can be approximated by evaluating the frequency of intensity values in the image. High entropy can reflect, for example, a large amount of anatomical detail associated with a clear image, or it can reflect, for example, a cloud of particulate eddies obscuring an anatomical site and will be associated with a non-clear image. A temporal course of entropy measurements can help to distinguish, for example, a clear image of high entropy from a non-clear image of high entropy. For example, a temporal course of entropy measurements of a series of images can distinguish between these two types of images by the variability of entropy in the non-clear image.

[0036] Alternatively, low entropy can reflect loss of contrast and / or obscuring of detail. Low entropy can also result from a very plain scene (e.g., a blank white wall). Image entropy metrics can be used to measure the amount of image entropy, including, for example, the total entropy in the image or the ratio of red entropy to cyan entropy in the image. In one embodiment, the ratio of red to cyan entropy can highlight the contribution of blood (represented by red entropy) relative to other fluids (represented by cyan entropy) to highlight the blood contribution. In some embodiments, the image entropy metrics can further encompass fluctuations in entropy in specific temporal bands in the image sequence. As described above, entropy fluctuations are measured over time to distinguish between clear and non-clear images. In particular, rapid changes in the observed image's high entropy metric values can reflect chaotic entropy associated with a cloud of particulate eddies, indicating that the observed image is a non-clear image.

[0037] Optical flow analysis is used to characterize the changing scene between images. Optical flow analysis generally identifies neighborhoods in a given image that can correspond to neighborhoods in a previous image and identifies differences in their locations. One algorithm for estimating such location displacements is the Farneback algorithm. This analysis provides information about stationary and moving objects in the field of view, as well as systematic field of view motion (i.e., translation, rotation, forward and backward movement of the camera).

[0038] Machine learning systems can also be used to characterize and classify video segments according to obscuration level and scene content. For example, machine learning methods such as neural networks, convolutional neural networks, optimizers, linear and / or logistic regression classifiers can be used to discover new image analysis and / or combinations of the above entropy metrics to characterize and classify video segments. The machine learning methods can generate a scalar estimate of turbidity, as well as probabilities of blood and particulate field of view.

[0039] In another embodiment, other metrics can be used to classify video segments, such as, for example, spatial and temporal frequency decomposition. Spatial frequency analysis can effectively measure the sharpness of an image to provide direct insight into the image definition. Images with a cloudy particulate view can exhibit a relatively high spatial frequency similar to that of images with high anatomical detail. Distinguishing the high spatial frequency of images with a cloudy view from images with high anatomical detail can be accomplished by temporal frequency analysis. As noted above, temporal frequency analysis refers to tracking of a metric over time.

[0040] The machine learning system can optimize the nonlinear combiner and various assessments, including the particulate / calculus assessment, blood assessment, and definition assessment, which will be described in detail below. Various metrics can be derived by linear combinations of the pixel values, intensity histogram values, nonlinear functions (e.g., logarithm), and temporal sequences of these values described above. The machine learning system can use algorithms to find alternative combinations of the values described above that do not directly correspond to spatial frequency or entropy, but more effectively correspond to the observer’s impression of turbidity.

[0041] The machine learning system operates on the feature space it is presented (i.e., the system values determine the classification). In one embodiment, the video machine learning system has video pixels directly input into the system. Image metrics (e.g., spatial spectrum) are computed from these values in common machine learning models. The feature space can be augmented by feed system analysis that can be helpful but not suitable for computation in the machine learning system. For example, entropy, which is thought to be useful for such computation, is not readily computed on many common video processing neural network structures. As a result, the machine learning system can perform better when provided with other analysis in addition to the raw pixels. Providing intensity histogram data (including its logarithm) to the neural network allows the network to incorporate entropy and entropy-like metrics, leading to faster convergence to better results. The system can recognize mathematical changes in the metrics as more optimal. Thus, careful enrichment of the feature space can significantly improve performance.

[0042] No particular metric derived directly from image data does a suitable job of estimating turbidity in a clinical setting under changing image backgrounds. The example embodiments describe methods for, for example, informing and tuning a neural network with entropy-related metrics to assess turbidity levels from image data.

[0043] The metrics and analysis described above can be input into a suitably designed machine learning algorithm that performs multiple assessments that can be output to a display to annotate the displayed image frames and / or can inform further analysis.

[0044] In a first example, a particle or stone assessment is performed. The particle / stone assessment determines characteristics of particles suspended in the fluid. Feature detection can be used to determine properties of the particles. For example, a kidney stone can be identified as a kidney stone and metrics such as stone size can be determined as well as a depth estimate (i.e., distance from the imager). Motion of the stone can also be estimated based in part on the optical flow analysis described above. Other particles can include, for example, proteins, tissue, protrusions, thrombus, etc. The particle / stone assessment can identify the particle, segment it in the image frame, and determine its size.

[0045] Particle identification and classification can be used to annotate the current image frame. For example, particles can be categorized and relevant metrics displayed to inform the operating physician and allow for faster decision making. The particle / stone analysis can also be used to manage the delivery of irrigation fluid by the fluid delivery mechanism 106. For example, while a generally cloudy view can indicate a need to increase fluid flow, the identification of a mobile stone can indicate a reduction in fluid flow to maintain the stone's position in the imager's view. Additionally, depending on the nature of the treatment, the particle / stone assessment can directly or indirectly affect the use of the treatment device 108. For example, in a lithotripsy procedure, the classification of a stone as, for example, too large to be retrieved, too large to pass naturally or too small to pass naturally, can automatically (or under physician direction) drive or affect the system's operation during the remainder of the intervention.

[0046] In a second example, a clarity assessment is performed. The clarity assessment determines a total measure of turbidity in the image. As described previously, a machine learning process with, for example, an image entropy metric as input can produce a new turbidity metric. The turbidity metric can be, for example, a scalar estimate of turbidity or some other metric developed by the machine learning algorithm. The clarity assessment can be used to annotate the currently displayed image and can also be used to manage the delivery of irrigation fluid.

[0047] In a third example, a blood assessment is performed. The blood assessment estimates a total measure of blood content within the cavity, for example parts per million (PPM), although other metrics can be used. The blood assessment can be used to annotate the currently displayed image and can also be used to manage the delivery of irrigation fluid. In addition, the blood assessment can inform, for example, a thrombus analysis, laser settings for, for example, a BPH procedure, etc.

[0048] In addition to the particle assessment, the clarity assessment, and the blood assessment discussed above, a fourth assessment can be used to manage fluid circulation in the imaged cavity within the organism. The fourth assessment relates to the type of intervention being performed during the endoscopic procedure. For example, the intervention can be a kidney stone treatment, a prostate resection, a uterine fibroid treatment, etc. A feature detector can be applied to the current image frame to determine the type of intervention activity being performed. For example, during a laser prostate resection, a laser device will be located within the FOV of the imager, and can be identified as a laser device by the feature detector.

[0049] In another example, a stone basket device can be identified by the feature detector, etc. Once identified, an assessment of the stage of the intervention is made. For example, the system can measure the degree to which a stone has been fragmented or otherwise reduced in size, and / or can determine when the stone is collected in the basket for retrieval, etc. Thus, during a laser lithotripsy procedure, the intervention assessment can be used in conjunction with, for example, at least the particle / stone assessment to identify the stage of the intervention that can affect the desired flow characteristics (e.g., field of view free of stone, field of view with stone, laser treatment in progress). The type of intervention can affect the optimal flow rate of the irrigation system fluid. For example, a laser prostate resection can require a higher flow rate than, for example, a kidney stone procedure due to the heating of the tissue. The heating of the tissue can result in a complementary heating of the fluid within the cavity. Faster fluid flow in and out can serve to maintain a more constant temperature within the cavity and prevent damage to healthy tissue.

[0050] Figure 2 A method 200 for managing irrigation fluid delivery in a closed loop feedback system in accordance with various example embodiments of the present application is shown. The method 200 can employ some or all of the metrics described above. Certain of the described metrics can be derived from a single image frame, while other metrics can be derived from a sequence of image frames. Thus, closed loop adjustment of fluid delivery can be performed based on multiple image frames. However, given the fast frame rate of imagers commonly used in the art, the fluid delivery adjustments are applied on a sufficiently fast basis such that the adjustments appear substantially continuous from the perspective of the operating physician.

[0051] In 205, a sequence of image frames is captured by the imager 104. As described above, the total number of images required for a given calculation can vary, and thus the number of images in the sequence can vary accordingly.

[0052] In 210, some or all of the metrics described above are determined from one or more of the image frames in the sequence. For example, the image entropy metric can be derived from each image and determined therefrom as the image is captured. In another example, the pixel metrics are determined directly from the image.

[0053] In 215, the various metrics and analyses are processed using a machine learning combiner algorithm / function to characterize and classify the image or video segment according to the level of occlusion and scene content. As noted above, the machine learning algorithm can combine the metrics in various ways and self-adjust the weight or use of the metrics according to the analyses discovered. It is noted that in addition to the treatment feature detection and intervention stage assessment, all of the above-described metrics / analyses are input to the machine learning combiner algorithm, whereas the treatment feature / phase detection / assessment is determined directly from the image frames and does not inform the particulate assessment, blood assessment, or clarity assessment.

[0054] In 220, the flow management algorithm processes the particulate assessment, blood assessment, clarity assessment, and intervention stage assessment to determine whether a fluid delivery adjustment of the irrigation fluid is required and, if so, to determine an adjustment value.

[0055] In 225, the adjustment value is fed back to the processor and the flow provided by the fluid delivery mechanism is adjusted.

[0056] The method steps 205-225 are performed on a substantially continuous basis, providing a closed loop feedback system for managing fluid flow during an endoscopic intervention.

[0057] In certain of the above-described metrics, particularly those related to the blood assessment and clarity assessment, can be used to algorithmically provide image enhancement to compensate for the blood-colored view in endoscopic interventions. The obscuring of features by blood can be managed by algorithmically subtracting blood features from the image and enhancing blood-obscured features. The methods 300 and 400 described below can be applied individually or in combination.

[0058] Figure 3 A method 300 for enhancing visibility through a liquid medium obscured by blood is shown in accordance with a first exemplary embodiment. The method 300 includes certain steps of the method 200 for managing irrigation fluid delivery (or those similar to the method 200) and can be used individually or in combination therewith.

[0059] In 305, a sequence of images is captured by the imager 104 in accordance with step 205 of the method 200. In 310, each image is divided into fixed size regions. Each region can be, for example, 20x20 pixels.

[0060] In 315, certain ones of the aforementioned metrics are computed and processed for each region using the aforementioned machine learning combiner algorithm, similar to steps 210-215 of method 200. In this way, the blood content of each region is assessed, and a "blood map" is generated to characterize each region in terms of the presence of blood. Specifically, the metrics informing the machine learning combiner algorithm for blood assessment are determined and processed to classify each region. No granule / calculus assessment or clarity assessment need be performed for method 300, or any interventional device features determined. However, in another embodiment, clarity assessment can be performed in the case, for example, where the hazy view is caused by particles other than blood.

[0061] In 320, the values of the blood map are multiplied by the regional frame-to-frame changes of each metric associated with blood detection in the combiner, as adjusted for optical flow. In this way, the individual image sub-regions can be leveraged to locate regions obscured by blood, and image enhancement is selectively or proportionally applied to mitigate the loss of image detail caused by blood.

[0062] In 325, the values produced from 320 are used to scale a convolution kernel or color transform computed to be complementary to each metric, which is then applied to its region and sigmoidally scaled at the region boundaries. The entropy-dependent kernel will use a low-pass filter kernel (for positive entropy dependence) or a high-pass filter kernel (for negative entropy dependence) applied to the relevant color channel.

[0063] In 330, the image frame is recombined from the spatial frequency representation of each region including the phase and the magnitude of the changes to reduce the importance of the blood color foreground and enhance the background of interest. Depending on the scaling of the convolution kernel, the visual effect of blood can be attenuated to varying degrees on the image. The enhancement can be applied to the magnitude component of the spatial frequency, which can then be recombined with its phase component and converted back to a regular image.

[0064] Figure 4 A method 400 for enhancing visibility through a liquid medium obscured by blood is shown, according to a second exemplary embodiment.

[0065] In 405, the current image frame is correlated with a super-resolution image map using a deformable non-linear transformation. A super-resolution image map can generally be defined as an improved resolution image generated from a fusion of lower resolution images. The current image can be fused with the super-resolution image map to enhance the resolution of the current image.

[0066] In 410, a "blood mask" is created based on the aforementioned metrics associated with blood detection, which reflects the spatial distribution of the inferred blood turbidity in a given image frame.

[0067] In 415, the current image frame is fused with the warped super-resolution map to generate an enhanced frame that is a weighted average of the current frame and the super-resolution map. When the blood mask value is low, the pixel is driven toward the pixel of the original image frame, and when the mask value is high, the pixel is driven toward the pixel of the super-resolution map.

[0068] Figure 5 A flowchart 500 is shown that describes a combined embodiment that includes elements of the methods 200, 300, and 400. Those skilled in the art will appreciate that the various metrics and analyses described previously can be used individually or in a variety of combinations.

[0069] In 505, images are extracted from the endoscope imaging feed. As described previously, certain metrics / analyses, such as, for example, optical flow analysis, require multiple images to determine. However, for the flowchart 500, it can be assumed that enough previous images have been captured to perform each calculation.

[0070] In 510, the images are separated into their red and cyan components for the calculation of various metrics. For example, in 515, red and cyan entropy values are calculated. Metrics such as, for example, a red to cyan entropy ratio can be calculated, or the red cyan values can be used directly. In 520, a spatial frequency analysis is performed on the red and cyan spectral components. In 525, the red / cyan entropy values and the spatial frequency analysis inform a temporal frequency analysis. Those skilled in the art will appreciate that a temporal frequency analysis requires a series of pictures to analyze the differences between them. For example, entropy fluctuations are measured over time. In 530, color balance is performed in the context of comparing the red and cyan components. In 535, a pixel change rate analysis is performed in which pixel values between images are compared. In 540, an optical flow analysis is performed to correlate consecutive images and derive information such as, for example, the motion of objects or tissue over time.

[0071] In 545, the above-described metrics / analyses are combined in various ways by a machine learning system to generate granular information (granule / stone assessment 550), blood content information (blood assessment 555), and turbidity information (clarity assessment 560). An intervention assessment is also performed in which intervention features in the images are identified (treatment device feature detection 565) and the stage of the intervention is assessed (treatment stage assessment 570). The granular, blood, clarity, and intervention analyses are used to control the fluid flow of the irrigation system 575. As described previously, the fluid flow management 575 can include adjusting the flow rate, pressure, or the manner in which the fluid is dispensed.

[0072] In 580, the current image can be enhanced by removing or mitigating the blood component of the image according to method 300 or 400. In 585, the current image can be annotated according to method 200. The current image can also be annotated with the blood or turbidity metric values derived in the blood / clarity assessment. In 590, the enhanced / annotated image is displayed. As more images are captured, the above steps / metrics / analysis are performed or determined.

[0073] In another embodiment, a computer readable medium includes instructions that, when executed by a computer, cause the computer to perform the various image processing steps / analysis described above (e.g., determine turbidity metrics).

[0074] Those skilled in the art will understand that variations can be made to the above-described embodiments without departing from the inventive concepts. It should also be understood that structural features and methods associated with one of the embodiments can be incorporated into other embodiments. Accordingly, it is to be understood that the application is not to be limited by the specific disclosed embodiments, which can be modified in various ways.

Claims

1. An endoscope system, comprising: An endoscopic imager configured to capture multiple image frames of a target site within a living body over a time span; and a processor configured to: determine one or more image metrics for each of the first and second image frames among the plurality of image frames captured within the time span; The one or more image metrics of the first image frame are compared with the one or more image metrics of the second image frame to determine the changes in the one or more image metrics over the time span; Based on the changes in the one or more image measures, a turbidity measure for the target region is determined; The processor identifies, classifies, and determines the properties of particles in the current image frame among multiple image frames; and controls a fluid delivery mechanism to provide fluid to a target area and adjusts the fluid flow rate according to the identified particles, wherein the processor reduces the fluid flow rate when the identified particles are kidney stones and increases the fluid flow rate when the identified particles are associated with turbid vision.

2. The endoscope system according to claim 1, wherein, The one or more image metrics are image entropy metrics that include red entropy metrics and cyan entropy metrics.

3. The endoscope system according to claim 2, wherein, The processor is also configured to: estimate the blood content in the current image frame; and modify the current image frame to reduce the visual effect of the blood content and enhance the rest of the current image frame.

4. The endoscope system according to claim 1, further comprising: A display configured to annotate the current image frame using the turbidity metric.

5. The endoscopic system according to claim 4, wherein, The display is also configured to: classify identified particles into a category; and annotate the current image frame using particle classification.

6. The endoscopic system according to claim 5, wherein, The identified particles are kidney stones, and the classification is related to the size of the kidney stones.

7. An endoscope system, comprising: An endoscopic imager configured to capture multiple image frames of a target site within a living body over a time span; A fluid delivery mechanism configured to deliver irrigation fluid to the target site to clarify the field of view of the endoscopic imager; and a processor configured to: track changes in one or more image metrics in the plurality of image frames by comparing one or more image metrics frame by frame over the time span; Based on the changes, determine the turbidity measure of at least one image frame among the plurality of image frames; Identify, classify, and determine the properties of particles in the current image frame across multiple image frames; Based on the turbidity measurement or the identified particles, a fluid delivery adjustment for the irrigation fluid is determined; based on the determined fluid delivery adjustment, the fluid delivery mechanism is controlled to adjust the fluid delivery provided by the fluid delivery mechanism, wherein when the identified particles are kidney stones, the processor reduces the fluid flow rate, and when the identified particles are associated with turbid vision, the processor increases the fluid flow rate.

8. The endoscope system according to claim 7, wherein, The processor is also configured to determine the type of intervention activity based on feature detection of the interventional instrument.

9. The endoscope system according to claim 8, wherein, The processor is also configured to: determine the stage of the intervention activity type; and adjust the fluid delivery provided by the fluid delivery mechanism based on the stage of the intervention activity type.

10. The endoscope system according to any one of claims 7-9, wherein, The processor is also configured to adjust the fluid delivery provided by the fluid delivery mechanism based on particle classification.

11. The endoscopic system according to claim 10, wherein, The particles are kidney stones, and the particle classification is related to the size of the kidney stones.

12. The endoscope system according to any one of claims 7-9, wherein, The processor is also configured to: determine a blood metric of the at least one image frame; and adjust the fluid delivery provided by the fluid delivery mechanism based on the blood metric.

13. The endoscope system according to claim 12, wherein, The processor is also configured to: determine an image entropy metric for the at least one image frame.

14. The endoscopic system according to claim 13, wherein, The turbidity metric and the blood metric are determined in part based on the image entropy metric.

Images