Hierarchical object detection in surgical images

Abstract

A computer-implemented method for performing a hierarchical object detection in surgical images includes providing a surgical image to a network including a backbone coupled to a bounding box head and a classification head. A hierarchical binary cross-entropy loss of an output of the classification head is determined, where the output of the classification head includes a multi-class probability vector with a hierarchical structure. The network is trained using the hierarchical binary cross-entropy loss to reason with respect to relationships between classes of the hierarchical structure. Class routing is performed at inference time to identify a label associated with a classification of the network for an object in a bounding box defined by the bounding box head.

Description

HIERARCHICAL OBJECT DETECTION IN SURGICAL IMAGESCROSS-REFERENCE TO RELATED APPLICATION

[0001] This application claims the benefit of U.S. Provisional Patent Application Serial No. 63 / 562,329, filed March 7, 2024, the entire content of which is incorporated herein by reference.BACKGROUND

[0002] The present disclosure relates in general to computing technology and relates more particularly to computing technology for the hierarchical object detection in surgical images.

[0003] Computer-assisted systems, particularly computer-assisted surgery systems (CASs), rely on video data digitally captured during surgery. Such video data can be stored and / or streamed. In some cases, video data can be augmented to highlight potential features of interest or detected events / conditions. Further, features within a surgical scene may be identified for tracking or other purposes without modifying the content of the video through overlays.

[0004] Object identification in surgical scenes may provide valuable information for real-time guidance and post-operative analysis of robotic-assisted laparoscopy.Unfortunately, however, object identification is challenging due to ambiguities caused by similar appearances of objects; occlusion by blood, visceral fat, and / or smoke, and camera pose. This can lead to missed detection or incorrect object identification.SUMMARY

[0005] According to an aspect, a computer-implemented method for performing hierarchical object detection in surgical images is provided. The method includes providing a surgical image to a network including a backbone coupled to a bounding box head and a classification head. A hierarchical binary cross-entropy loss of an output ofthe classification head is determined, where the output of the classification head includes a multi-class probability vector with a hierarchical structure. The network is trained using the hierarchical binary cross-entropy loss to reason with respect to relationships between classes of the hierarchical structure. Class routing is performed at inference time to identify a label associated with a classification of the network for an object in a bounding box defined by the bounding box head.

[0006] According to another aspect, a system includes a data store including video data associated with a surgical procedure and a machine learning training system for training a network using surgical images extracted from the video data. The system is configured to provide the surgical images to the network including a backbone coupled to a bounding box head and a classification head. The system is further configured to determine a hierarchical binary cross-entropy loss of an output of the classification head, where the output of the classification head includes a multi-class probability vector with a hierarchical structure. The system is also configured to train the network using the hierarchical binary cross-entropy loss to reason with respect to relationships between classes of the hierarchical structure.

[0007] According to an aspect, a computer program product is provided and includes a memory device having computer executable instructions stored thereon, which when executed by one or more processors cause the one or more processors to perform a plurality of operations for performing hierarchical object detection in surgical images. The plurality of operations include providing a surgical image to a network including a backbone coupled to a classification head, determining a hierarchical binary crossentropy loss of an output of the classification head, where the output of the classification head includes a multi-class probability vector with a hierarchical structure, training the network using the hierarchical binary cross-entropy loss to reason with respect to relationships between classes of the hierarchical structure, and performing class routing at inference time to identify a label associated with a classification of the network for an object.

[0008] The above features and advantages, and other features and advantages, of the disclosure are readily apparent from the following detailed description when taken in connection with the accompanying drawings.BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The specifics of the exclusive rights described herein are particularly pointed out and distinctly claimed in the claims at the conclusion of the specification. The foregoing and other features and advantages of the aspects of the disclosure are apparent from the following detailed description taken in conjunction with the accompanying drawings in which:

[0010] FIG. 1 depicts a computer-assisted surgery (CAS) system according to one or more aspects;

[0011] FIG. 2 depicts a surgical procedure system according to one or more aspects;

[0012] FIG. 3 depicts a system for analyzing video and data according to one or more aspects;

[0013] FIG. 4 depicts a flow diagram for hierarchical detection, according to one or more aspects;

[0014] FIG. 5 depicts a hierarchical binary cross-entropy algorithm, according to one or more aspects;

[0015] FIG. 6 depicts a class routing algorithm, according to one or more aspects;

[0016] FIG. 7 A depicts a surgical image with bounding boxes identifying detected objects, according to one or more aspects;

[0017] FIG. 7B depicts a hierarchical representation of surgical instrument classes with respect to the surgical image of FIG. 7 A, according to one or more aspects;

[0018] FIG. 8 depicts a flowchart of a method of performing hierarchical object detection in surgical images, according to one or more aspects; and

[0019] FIG. 9 depicts a block diagram of a computer system, according to one or more aspects.

[0020] The diagrams depicted herein are illustrative. There can be many variations to the diagrams and / or the operations described herein without departing from the spirit of the described aspects. For instance, the actions can be performed in a differing order, or actions can be added, deleted, or modified. Also, the term “coupled” and variations thereof describe having a communications path between two elements and do not imply a direct connection between the elements with no intervening elements / connections between them. All of these variations are considered a part of the specification.DETAILED DESCRIPTION

[0021] Exemplary aspects of the technical solutions described herein include systems and methods for the hierarchical object detection in surgical images.

[0022] In order to improve surgical video and image analysis, aspects of technical solutions are described herein and may use the hierarchical nature of classes in an object detection task for surgical instrument detection in laparoscopic surgery images extracted from surgical video.

[0023] Technical solutions are described herein to address such technical challenges. Particularly, technical solutions herein may facilitate improved detection accuracy of objects, such as surgical instruments, in surgical images. The improved detection accuracy can be beneficial for multiple aspects of real-time and post-operative analytics, including, for example, advancements in computer-assisted intra-operative guidance.

[0024] Rob otic- Assisted Surgery (RAS), such as robotic laparoscopic surgery, has enabled surgeons to perform minimally invasive procedures with greater precision,thereby resulting in less pain, scarring, reduced blood loss, and faster recovery. A key component of RAS which allows for increased surgical precision is visual feedback via integrated imaging and display technology. This technology is typically capable of providing the surgeon with a high resolution, magnified view of the internal anatomy of interest and the surgical tools being used. Furthermore, solutions have been developed to allow post-operative analysis of recorded surgical video. However, interpretation of surgical video can be challenging due to several reasons, including but not limited to, occlusion caused by blood, visceral fat, inflammation, smoke (e.g., from electrocautery), as well as other anatomical structures that are not the target structures of interest, reduced reference due to magnification / camera angle, and specularity.

[0025] Robot-assisted surgery allows surgeons to perform minimally invasive procedures with increased precision. In contrast to open surgery, this technique involves small incisions, usually under 5 mm, resulting in less scarring and faster patient recovery. During the procedure, the surgeon is presented with a digital magnified live feed of a narrow operating field of view, in which surgical instruments are inserted via small tissue openings. In this context, context-aware assistance systems aim at extracting clinically relevant information from the image during the procedure for human-machine collaboration. Detection and classification of instruments present in the field of view is a necessary step towards computer-assisted intervention (CAI) in the next generation of operating room (ORs) environments. Downstream tasks can consist of surgical phase classification, instrument out of view warnings, or key-points pose estimation, amongst others.

[0026] The spectrum of instruments involved in laparoscopic surgery is large, and is often organized in a structured manner. For instance, passive (or no-energy) instruments, such as forceps or retractors, are instruments that do not deliver any type of energy to surrounding tissue. Active (or energy) instruments, such as monopolar shears or bipolar fenestrated tools, deliver heat or can cauterize surrounding tissue. In many situations,some instruments share design similarities in some part. It is common that two instruments may be indistinguishable when partly occluded.

[0027] Deep learning methods for object detection and pose estimation is an active area of research. State of the art methods can use a large pre-trained Convolutional Neural Network (CNN) backbone followed with shallow task specific head branches. The “Yolo” network family can be helpful in the field of object detection and pose estimation. Yolo-based networks have two main advantages. Firstly, they belong to the group of one-stage detectors, as opposed to two-stage detectors such as Faster RCNN approaches, and are therefore simpler by design. Secondly, they benefit from a fast inference speed. In the Yolo network family, one approach is referred to as “Yolov8” and presents a compromise between inference speed and accuracy. Additionally to incorporating a more efficient backbone, Yolov8 also includes a spatial attention mechanism inspired from Cross Stage Partial Networks.

[0028] According to aspects disclosed herein, a hierarchy-aware loss, referred to as “hBCE”, can be defined for a classification branch of a network. A heuristic-based approach can be used to isolate an output label at inference time. One or more hierarchical detection metrics can be used to evaluate the results. The use of hBCE in the classification branch can result in a significant increase in detection precision and recall. Using a hierarchy-aware metric for a detection task may be more appropriate in situations where the input classes show high taxonomy.

[0029] According to aspects, the hierarchical nature of surgical instruments can be leveraged for the task of object bounding box detection in a surgical image. At data preprocessing time, pseudo-labels can be introduced to transform standard hierarchyagnostic labels into a directed graph representing a hierarchy of classes. Objects can be re-categorized depending on the visibility and occlusion of the instrument in the frame. At train time, a hierarchy-aware loss can be introduced to replace classical one-hot binary cross entropy losses of detection networks. At inference / test time, a heuristic-based routing algorithm can be used to define an inference single-label decision.

[0030] Turning now to FIG. 1, an example computer-assisted system (CAS) system 100 is generally shown in accordance with one or more aspects. The CAS system 100 includes at least a computing system 102, a video recording system 104, and a surgical instrumentation system 106. As illustrated in FIG. 1, an actor 112 can be medical personnel that uses the CAS system 100 to perform a surgical procedure on a patient 110. Actor 112 may be any medical personnel such as a surgeon, assistant, nurse, administrator, or any other actor that interacts with the CAS system 100 in a surgical environment. The surgical procedure can be any type of surgery, such as but not limited to cataract surgery, laparoscopic cholecystectomy, endoscopic endonasal transsphenoidal approach (eTSA) to resection of pituitary adenomas, or any other surgical procedure. In other examples, actor 112 can be a technician, an administrator, an engineer, or any other such personnel that interacts with the CAS system 100. For example, actor 112 can record data from the CAS system 100, configure / update one or more attributes of the CAS system 100, review past performance of the CAS system 100, repair the CAS system 100, and / or the like including combinations and / or multiples thereof.

[0031] A surgical procedure can include multiple phases, and each phase can include one or more surgical actions. A “surgical action” can include an incision, a compression, a stapling, a clipping, a suturing, a cauterization, a sealing, or any other such actions performed to complete a phase in the surgical procedure. A “phase” represents a surgical event that is composed of a series of steps (e.g., closure). A “step” refers to the completion of a named surgical objective (e.g., hemostasis). During each step, certain surgical instruments 108 (e.g., forceps) are used to achieve a specific objective by performing one or more surgical actions. In addition, a particular anatomical structure of the patient may be the target of the surgical action(s).

[0032] The video recording system 104 includes one or more cameras 105, such as operating room cameras, endoscopic cameras, laparoscopic cameras, and / or the like including combinations and / or multiples thereof. The cameras 105 capture video data of the surgical procedure being performed. The video recording system 104 includes one ormore video capture devices that can include cameras 105 placed in the surgical room to capture events surrounding (i.e., outside) the patient being operated upon. The video recording system 104 further includes cameras 105 that are passed inside (e.g., endoscopic cameras) the patient 110 to capture endoscopic data. The endoscopic data provides video and images of the surgical procedure.

[0033] Computing system 102 includes one or more memory devices, one or more processors and a user interface device, among other components. All or a portion of the computing system 102 shown in FIG. 1 can be implemented, for example, by all or a portion of computer system 800 of FIG. 9. Computing system 102 can execute one or more computer-executable instructions. The execution of the instructions facilitates the computing system 102 to perform one or more methods, including those described herein. Computing system 102 can communicate with other computing systems via a wired and / or a wireless network. In one or more examples, the computing system 102 includes one or more trained machine learning models that can detect and / or predict features of / from the surgical procedure that is being performed or has been performed earlier.Features can include structures, such as anatomical structures, surgical instruments 108 in the captured video of the surgical procedure. Features can further include events, such as phases and / or actions in the surgical procedure. Features that are detected can further include the actor 112 and / or patient 110. Based on the detection, the computing system 102, in one or more examples, can provide recommendations for subsequent actions to be taken by the actor 112. Alternatively, or in addition, the computing system 102 can provide one or more reports based on the detections. The detections by the machine learning models can be performed in an autonomous or semi-autonomous manner.

[0034] Machine learning models can include artificial neural networks, such as deep neural networks, convolutional neural networks, recurrent neural networks, vision transformers, encoders, decoders, or any other type of machine learning model. Machine learning models can be trained in a supervised, unsupervised, or hybrid manner. The machine learning models can be trained to perform detection and / or prediction using oneor more types of data acquired by the CAS system 100. For example, machine learning models can use the video data captured via the video recording system 104. Alternatively, or in addition, the machine learning models use the surgical instrumentation data from the surgical instrumentation system 106. In yet other examples, the machine learning models use a combination of video data and surgical instrumentation data.

[0035] Additionally, in some examples, the machine learning models can also use audio data captured during the surgical procedure. The audio data can include sounds emitted by the surgical instrumentation system 106 while activating one or more surgical instruments 108. Alternatively, or in addition, the audio data can include voice commands, snippets, or dialog from one or more actors 112. The audio data can further include sounds made by the surgical instruments 108 during their use.

[0036] In one or more examples, the machine learning models can detect surgical actions, surgical phases, anatomical structures, surgical instruments, and various other features from the data associated with a surgical procedure. The detection can be performed in real-time in some examples. Alternatively, or in addition, computing system 102 analyzes the surgical data, i.e., the various types of data captured during the surgical procedure, in an offline manner (e.g., post-surgery). In one or more examples, the machine learning models detect surgical phases based on detecting some of the features, such as the anatomical structure, surgical instruments, and / or the like including combinations and / or multiples thereof.

[0037] A data collection system 150 can be employed to store the surgical data, including the video(s) captured during the surgical procedures. The data collection system 150 includes one or more storage devices 152. The data collection system 150 can be a local storage system, a cloud-based storage system, or a combination thereof. Further, the data collection system 150 can use any type of cloud-based storage architecture, for example, public cloud, private cloud, hybrid cloud, and / or the like including combinations and / or multiples thereof. In some examples, the data collection system can use a distributed storage, i.e., storage devices 152 are located at differentgeographic locations. The storage devices 152 can include any type of electronic data storage media used for recording machine-readable data, such as semiconductor-based, magnetic-based, optical-based storage media, and / or the like including combinations and / or multiples thereof. For example, the data storage media can include flash-based solid-state drives (SSDs), magnetic-based hard disk drives, magnetic tape, optical discs, and / or the like including combinations and / or multiples thereof.

[0038] In one or more examples, the data collection system 150 can be part of the video recording system 104, or vice-versa. In some examples, the data collection system 150, the video recording system 104, and the computing system 102, can communicate with each other via a communication network, which can be wired, wireless, or a combination thereof. The communication between the systems can include the transfer of data (e.g., video data, instrumentation data, and / or the like including combinations and / or multiples thereof), data manipulation commands (e.g., browse, copy, paste, move, delete, create, compress, and / or the like including combinations and / or multiples thereof), data manipulation results, and / or the like including combinations and / or multiples thereof. In one or more examples, the computing system 102 can manipulate the data already stored / being stored in the data collection system 150 based on outputs from the one or more machine learning models (e.g., phase detection, anatomical structure detection, surgical tool detection, and / or the like including combinations and / or multiples thereof). Alternatively, or in addition, the computing system 102 can manipulate the data already stored / being stored in the data collection system 150 based on information from the surgical instrumentation system 106.

[0039] In one or more examples, the video captured by the video recording system 104 is stored on the data collection system 150. In some examples, the computing system 102 curates parts of the video data being stored on the data collection system 150. In some examples, the computing system 102 filters the video captured by the video recording system 104 before it is stored on the data collection system 150. Alternatively, or inaddition, the computing system 102 filters the video captured by the video recording system 104 after it is stored on the data collection system 150.

[0040] Turning now to FIG. 2, a surgical procedure system 200 is generally shown according to one or more aspects. The example of FIG. 2 depicts a surgical procedure support system 202 that can include or may be coupled to the CAS system 100 of FIG. 1. The surgical procedure support system 202 can acquire image or video data using one or more cameras 204. The surgical procedure support system 202 can also interface with one or more sensors 206 and / or one or more effectors 208. The sensors 206 may be associated with surgical support equipment and / or patient monitoring. The effectors 208 can be robotic components or other equipment controllable through the surgical procedure support system 202. The surgical procedure support system 202 can also interact with one or more user interfaces 210, such as various input and / or output devices. The surgical procedure support system 202 can store, access, and / or update surgical data 214 associated with a training dataset and / or live data as a surgical procedure is being performed on patient 110 of FIG. 1. The surgical procedure support system 202 can store, access, and / or update surgical objectives 216 to assist in training and guidance for one or more surgical procedures. User configurations 218 can track and store user preferences.

[0041] Turning now to FIG. 3, a system 300 for analyzing video and data is generally shown according to one or more aspects. In accordance with aspects, the video and data is captured from video recording system 104 of FIG. 1. The analysis can result in predicting features that include surgical phases and structures (e.g., instruments, anatomical structures, and / or the like including combinations and / or multiples thereof) in the video data using machine learning. System 300 can be the computing system 102 of FIG. 1, or a part thereof in one or more examples. System 300 uses data streams in the surgical data to identify procedural states according to some aspects.

[0042] System 300 includes a data reception system 305 that collects surgical data, including the video data and surgical instrumentation data. The data reception system305 can include one or more devices (e.g., one or more user devices and / or servers) located within and / or associated with a surgical operating room and / or control center. The data reception system 305 can receive surgical data in real-time, i.e., as the surgical procedure is being performed. Alternatively, or in addition, the data reception system 305 can receive or access surgical data in an offline manner, for example, by accessing data that is stored in the data collection system 150 of FIG. 1.

[0043] System 300 further includes a machine learning processing system 310 that processes the surgical data using one or more machine learning models to identify one or more features, such as surgical phase, instrument, anatomical structure, and / or the like including combinations and / or multiples thereof, in the surgical data. It will be appreciated that machine learning processing system 310 can include one or more devices (e.g., one or more servers), each of which can be configured to include part or all of one or more of the depicted components of the machine learning processing system 310. In some instances, a part or all of the machine learning processing system 310 is cloudbased and / or remote from an operating room and / or physical location corresponding to a part or all of data reception system 305. It should be appreciated that several components of the machine learning processing system 310 are depicted and described herein. However, the components are just one example structure of the machine learning processing system 310, and that in other examples, the machine learning processing system 310 can be structured using a different combination of the components. Such variations in the combination of the components are encompassed by the technical solutions described herein.

[0044] The machine learning processing system 310 includes a machine learning training system 325, which can be a separate device (e.g., server) that stores its output as one or more trained machine learning models 330. The machine learning models 330 are accessible by a machine learning execution system 340. The machine learning execution system 340 can be separate from the machine learning training system 325 in some examples. In other words, in some aspects, devices that “train” the models are separatefrom devices that “infer,” i.e., perform real-time processing of surgical data using the trained machine learning models 330.

[0045] Machine learning processing system 310, in some examples, further includes a data generator 315 to generate simulated surgical data, such as a set of synthetic images and / or synthetic video, in combination with real image and video data from the video recording system 104, to generate trained machine learning models 330. Data generator 315 can access (read / write) a data store 320 to record data, including multiple images and / or multiple videos. The images and / or videos can include images and / or videos collected during one or more procedures (e.g., one or more surgical procedures). For example, the images and / or video may have been collected by a user device worn by the actor 112 of FIG. 1 (e.g., surgeon, surgical nurse, anesthesiologist, and / or the like including combinations and / or multiples thereof) during the surgery, a non-wearable imaging device located within an operating room, an endoscopic camera inserted inside the patient 110 of FIG. 1, and / or the like including combinations and / or multiples thereof. The data store 320 is separate from the data collection system 150 of FIG. 1 in some examples. In other examples, the data store 320 is part of the data collection system 150.

[0046] Each of the images and / or videos recorded in the data store 320 for performing training (e.g., generating the machine learning models 330) can be defined as a base image and can be associated with other data that characterizes an associated procedure and / or rendering specifications. For example, the other data can identify a type of procedure, a location of a procedure, one or more people involved in performing the procedure, surgical objectives, and / or an outcome of the procedure. Alternatively, or in addition, the other data can indicate a stage of the procedure with which the image or video corresponds, rendering specification with which the image or video corresponds and / or a type of imaging device that captured the image or video (e.g., and / or, if the device is a wearable device, a role of a particular person wearing the device, and / or the like including combinations and / or multiples thereof). Further, the other data can include image-segmentation data that identifies and / or characterizes one or more objects (e.g.,tools, anatomical objects, and / or the like including combinations and / or multiples thereof) that are depicted in the image or video. The characterization can indicate the position, orientation, or pose of the object in the image. For example, the characterization can indicate a set of pixels that correspond to the object and / or a state of the object resulting from a past or current user handling. Localization can be performed using a variety of techniques for identifying objects in one or more coordinate systems.

[0047] The machine learning training system 325 uses the recorded data in the data store 320, which can include the simulated surgical data (e.g., set of synthetic images and / or synthetic video) and / or actual surgical data to generate the trained machine learning models 330. The trained machine learning models 330 can be defined based on a type of model and a set of hyperparameters (e.g., defined based on input from a client device). The trained machine learning models 330 can be configured based on a set of parameters that can be dynamically defined based on (e.g., continuous or repeated) training (i.e., learning, parameter tuning). Machine learning training system 325 can use one or more optimization algorithms to define the set of parameters to minimize or maximize one or more loss functions. The set of (learned) parameters can be stored as part of the trained machine learning models 330 using a specific data structure for a particular trained machine learning model of the trained machine learning models 330. The data structure can also include one or more non-learnable variables (e.g., hyperparameters and / or model definitions).

[0048] Machine learning execution system 340 can access the data structure(s) of the trained machine learning models 330 and accordingly configure the trained machine learning models 330 for inference (e.g., prediction, classification, and / or the like including combinations and / or multiples thereof). The trained machine learning models 330 can include, for example, a fully convolutional network adaptation, an adversarial network model, an encoder, a decoder, or other types of machine learning models. The type of the trained machine learning models 330 can be indicated in the correspondingdata structures. The trained machine learning models 330 can be configured in accordance with one or more hyperparameters and the set of learned parameters.

[0049] The trained machine learning models 330, during execution, receive, as input, surgical data to be processed and subsequently generate one or more inferences according to the training. For example, the video data captured by the video recording system 104 of FIG. 1 can include data streams (e.g., an array of intensity, depth, and / or RGB values) for a single image or for each of a set of frames (e.g., including multiple images or an image with sequencing data) representing a temporal window of fixed or variable length in a video. The video data that is captured by the video recording system 104 can be received by the data reception system 305, which can include one or more devices located within an operating room where the surgical procedure is being performed. Alternatively, the data reception system 305 can include devices that are located remotely, to which the captured video data is streamed live during the performance of the surgical procedure. Alternatively, or in addition, the data reception system 305 accesses the data in an offline manner from the data collection system 150 or from any other data source (e.g., local or remote storage device).

[0050] The data reception system 305 can process the video and / or data received. The processing can include decoding when a video stream is received in an encoded format such that data for a sequence of images can be extracted and processed. The data reception system 305 can also process other types of data included in the input surgical data. For example, the surgical data can include additional data streams, such as audio data, RFID data, textual data, measurements from one or more surgical instrum ents / sensors, and / or the like including combinations and / or multiples thereof, that can represent stimuli / procedural states from the operating room. The data reception system 305 synchronizes the different inputs from the different devices / sensors before inputting them in the machine learning processing system 310.

[0051] The trained machine learning models 330, once trained, can analyze the input surgical data, and in one or more aspects, predict and / or characterize features (e.g.,structures) included in the video data included with the surgical data. The video data can include sequential images and / or encoded video data (e.g., using digital video file / stream formats and / or codecs, such as MP4, MOV, AVI, WEBM, AVCHD, OGG, and / or the like including combinations and / or multiples thereof). The prediction and / or characterization of the features can include segmenting the video data or predicting the localization of the structures with a probabilistic heatmap. In some instances, the one or more trained machine learning models 330 include or are associated with a preprocessing or augmentation (e.g., intensity normalization, resizing, cropping, and / or the like including combinations and / or multiples thereof) that is performed prior to segmenting the video data. An output of the one or more trained machine learning models 330 can include image-segmentation or probabilistic heatmap data that indicates which (if any) of a defined set of structures are predicted within the video data, a location and / or position and / or pose of the structure(s) within the video data, and / or state of the structure(s). The location can be a set of coordinates in an image / frame in the video data. For example, the coordinates can provide a bounding box. The coordinates can provide boundaries that surround the structure(s) being predicted. The trained machine learning models 330, in one or more examples, are trained to perform higher-level predictions and tracking, such as predicting a phase of a surgical procedure and tracking one or more surgical instruments used in the surgical procedure.

[0052] While some techniques for predicting a surgical phase (“phase”) in the surgical procedure are described herein, it should be understood that any other technique for phase prediction can be used without affecting the aspects of the technical solutions described herein. In some examples, the machine learning processing system 310 includes a detector 350 that uses the trained machine learning models 330 to identify various items or states within the surgical procedure (“procedure”). The detector 350 can use a particular procedural tracking data structure 355 from a list of procedural tracking data structures. The detector 350 can select the procedural tracking data structure 355 based on the type of surgical procedure that is being performed. In one or more examples, the type of surgical procedure can be predetermined or input by actor 112. For instance, theprocedural tracking data structure 355 can identify a set of potential phases that can correspond to a part of the specific type of procedure as “phase predictions”, where the detector 350 is a phase detector.

[0053] In some examples, the procedural tracking data structure 355 can be a graph that includes a set of nodes and a set of edges, with each node corresponding to a potential phase. The edges can provide directional connections between nodes that indicate (via the direction) an expected order during which the phases will be encountered throughout an iteration of the procedure. The procedural tracking data structure 355 may include one or more branching nodes that feed to multiple next nodes and / or can include one or more points of divergence and / or convergence between the nodes. In some instances, a phase indicates a procedural action (e.g., surgical action) that is being performed or has been performed and / or indicates a combination of actions that have been performed. In some instances, a phase relates to a biological state of a patient undergoing a surgical procedure. For example, the biological state can indicate a complication (e.g., blood clots, clogged arteries / veins, and / or the like including combinations and / or multiples thereof), pre-condition (e.g., lesions, polyps, and / or the like including combinations and / or multiples thereof). In some examples, the trained machine learning models 330 are trained to detect an “abnormal condition,” such as hemorrhaging, arrhythmias, blood vessel abnormality, and / or the like including combinations and / or multiples thereof.

[0054] Each node within the procedural tracking data structure 355 can identify one or more characteristics of the phase corresponding to that node. The characteristics can include visual characteristics. In some instances, the node identifies one or more tools that are typically in use or available for use (e.g., on a tool tray) during the phase. The node also identifies one or more roles of people who are typically performing a surgical task, a typical type of movement (e.g., of a hand or tool), and / or the like including combinations and / or multiples thereof. Thus, detector 350 can use the segmented data generated by machine learning execution system 340 that indicates the presence and / orcharacteristics of particular objects within a field of view to identify an estimated node to which the real image data corresponds. Identification of the node (i.e., phase) can further be based upon previously detected phases for a given procedural iteration and / or other detected input (e.g., verbal audio data that includes person-to-person requests or comments, explicit identifications of a current or past phase, information requests, and / or the like including combinations and / or multiples thereof).

[0055] The detector 350 can output predictions, such as a phase prediction associated with a portion of the video data that is analyzed by the machine learning processing system 310. The phase prediction is associated with the portion of the video data by identifying a start time and an end time of the portion of the video that is analyzed by the machine learning execution system 340. The phase prediction that is output can include segments of the video where each segment corresponds to and includes an identity of a surgical phase as detected by the detector 350 based on the output of the machine learning execution system 340. Further, the phase prediction, in one or more examples, can include additional data dimensions, such as, but not limited to, identities of the structures (e.g., instrument, anatomy, and / or the like including combinations and / or multiples thereof) that are identified by the machine learning execution system 340 in the portion of the video that is analyzed. The phase prediction can also include a confidence score of the prediction. Other examples can include various other types of information in the phase prediction that is output. Further, other types of outputs of the detector 350 can include state information or other information used to generate audio output, visual output, and / or commands. For instance, the output can trigger an alert, an augmented visualization, identify a predicted current condition, identify a predicted future condition, command control of equipment, and / or result in other such data / commands being transmitted to a support system component, e.g., through surgical procedure support system 202 of FIG. 2.

[0056] It should be noted that although some of the drawings depict endoscopic videos being analyzed, the technical solutions described herein can be applied to analyze videoand image data captured by cameras that are not endoscopic (i.e., cameras external to the patient’s body) when performing open surgeries (i.e., not laparoscopic surgeries). For example, the video and image data can be captured by cameras that are mounted on one or more personnel in the operating room (e.g., surgeon). Alternatively, or in addition, the cameras can be mounted on surgical instruments, walls, or other locations in the operating room. Alternatively, or in addition, the video can be images captured by other imaging modalities, such as ultrasound.

[0057] Turning now to FIG. 4, a flow diagram 400 for hierarchical detection is depicted according to one or more aspects. It is noted that classes can be defined as a hierarchical structure in the form of an acyclic directed graph G(V,E), where nodes V are connected by directed edges E. A level “1” can be assigned to each node as the number of ancestors of that node. A hierarchy in data can have various origins. In the context of surgical instruments, for instance, a natural hierarchical structure can be defined from instrument design and occlusions. Formally, let A and B be two distinct instruments with similar designs until the joint-shaft, and only differing by the design of the end-effector. A natural hierarchy can be formed by introducing a common ancestor O to both nodes A and B. An instance of instrument A (resp. B) can therefore be re-labelled / re-categorized as O if endeffector key-points are not visible in a frame (e.g., a surgical image of a surgical video).

[0058] As shown, FIG. 4 depicts that surgical images 402 can be red-green-blue (RGB) frames from a laparoscopic surgical procedure video which are fed to a CNN backbone and Feature Pyramidal Network (FPN) 404, also referred to as backbone 404. As one example, a Yolov8 backbone can be used. Each output feature vector of the FPN 404 can then be fed to 3 separate heads to extract (1) a normalized bounding box by a bounding box head 406, (2) a set of normalized key-point positions by a pose head 408, and (3) a multi-class probability vector by a classification head 410. The combination of the FPN 404 and heads 406, 408, 410 may be referred to as network 405, which can be one of the machine learning models in the trained machine learning models 330 of FIG. 3 after training. At training time, an output prediction of the classification head 410 can bedirected to a hierarchy-aware binary cross-entropy loss, named “hBCE” 412, instead of using a single-label multi-class head.

[0059] As one example, to map a target one-hot class vector to a multi-hot vector: (1) ones can be assigned to all ancestor nodes Va, and (2) l / |Vs| can be assigned to each of the direct successor nodes Vs. This loss modification implies that the network 405 will reason with respect to node position in a graph. This is illustrated in a hierarchical binary crossentropy algorithm 500 of FIG. 5. A simple loss modification in the hierarchical binary cross-entropy algorithm 500 can push the network 405 to reason with respect to the relationship between classes. Higher loss values can affect scenarios where an instance is misclassified with a node further away in the hierarchy.

[0060] In some aspects, an immediate consequence of the loss modification described above is that the output of the classification head 410 may not be suitable for performing a naive argmax to recover the output label of an instance. Class routing 414 can be used to account for classification issues using a class routing algorithm, such as class routing algorithm 600 of FIG. 6.

[0061] Given the hierarchical framework as previously described, it follows that the class routing algorithm 600 is configured to take the hierarchical graph into account. In other words, the decision over the final instance label is made with respect to the class relationships imposed by G. The class routing algorithm 600 can include the following properties:• A node should only be chosen if its prediction score and all prediction scores of its ancestors are high.• A node should only be chosen if there is no ambiguity within the nodes of that level.• Provided multiple nodes conform with the above requirements, the node at the highest level should be chosen.

[0062] FIG. 7A depicts a surgical image 650 with bounding boxes identifying detected objects, according to one or more aspects. FIG. 7B depicts a hierarchical representation 680 of surgical instrument classes with respect to the surgical image 650 of FIG. 7A, according to one or more aspects. The hierarchical approach, as described herein, can be generalized and applied to a variety of detection and pose estimation tasks where the data is hierarchical. The methodology can be evaluated, for example, in the context of surgical instrument detection in laparoscopic video frames. In laparoscopic surgical procedures, a monopolar curved shears (monop) can be the main active instrument used. Alongside the monop, two major energy instruments can be used, such as a fenestrated bipolar grasper (fenes) and / or a maryland bipolar forceps (maryl), as illustrated in the example of FIG. 7A. These two last instruments can have substantially similar designs up until the end-effector joint. Therefore, the instruments may be indistinguishable from each other unless a tip of at least one of the instruments is visible. Furthermore, other passive instruments (noeng) can have similar designs for the shaft part of the instrument, and therefore can appear indistinguishable from energy graspers (enrgg) if only the shaft is visible. These design similarities drive a hierarchical structure of surgical instruments as illustrated in the example of FIG. 7B. The instruments can be re-labeled from leaf nodes (i.e. fenes, maryl or noeng) according to the visibility of their end-effector and tip key-points. For instance, an instrument originally labelled as “fenes” can be re-labelled as “enrgg” if neither of the end-effector tips are visible, or as “grasp” if additionally, the joint-shaft is also not visible. Levels can be defined using heuristics on the visibility of the end-effector key-points. Example key -points can include a shaft start, shaft end, shaft joint, joint tip, and tips.

[0063] As a further illustrative example, experiments were performed using an internal dataset including about 340 videos of laparoscopic surgical procedures from about 12 different hospitals and 32 different surgeons. A subset of about 19,293 frames was randomly selected from the video feeds, and annotated in the following manner: (1) instruments present in the field of view were located via a bounding box and classified as described in FIG. 7A. Additionally, 6 key-points were identified and located for each annotated instrument, from shaft to end-effector. The dataset was randomly split into 3folds, train, val, and test at proportions 70%, 10% and 20% respectively. An FPN backbone was pretrained on a dataset for all our models, and the whole model was fine-tuned using the surgical instrument dataset for 100 epochs. An optimizer was used with a learning rate y = le - 4, and a weight decay 01 = 02 = 5e - 4. In addition to the hBCE loss as previously described, a complete distance intersection over union (loU) loss and a generalized focal loss were used for improved detection. A pose head can be fed to a simple mean squared error (MSE) loss and normalized by an instrument bounding box area. Optimal transport assignment (OTA) can be used for label assignment in the box regression loss. Various augmentation schemes can be applied to improve generalizations, such as cropping, rotations, and shear, exposure, temperature, blur, smoke, and horizontal and vertical flips. Typical metrics for classification and detection tasks can be used, such as flat precision, recall and F-score. These metrics may not be adequate when applied to hierarchical data. More formally, a metric may be suited to hierarchical data if (1) it punishes misclassifications between nodes that are further apart in the hierarchical graph, and (2) it punishes mis-classifications of a node with its siblings at a similar level than a node and its ancestor in the hierarchy. The following definitions follow these requirements:Where “c” denotes the class the metric is calculated on, Sxydenotes the sum over each paired ground truth class “x” and predicted class “y”, 6 denotes the Kroenecker delta function, and “a(i, j)” denotes the affinity between node “i” and node “j”. Note that these metrics are a generalization of a flat variant. In other words, the flat variants of Pnand Rocan be obtained by removing all edges in graph G. Unless specified otherwise, the above formulation can be used for evaluating detection accuracy.

[0064] Pose estimation can be evaluated using a PoseAcc metric that quantifies the accuracy of key-point estimations in a given pose estimation model. This can be computed as a ratio of correctly detected key-points (within a specified threshold in pixels) to the total number of annotated key-points. PoseAcc distinguishes between true positives (TP) and true negatives (TN). TP represents accurately detected and correctly classified keypoints, while TN corresponds to non-detected key-points that are genuinely absent. This metric can be assessed on a per-key-point or per-instrument basis, and can be further averaged across key-point types or instrument types, providing a comprehensive evaluation of pose estimation performance.

[0065] Table 1 illustrates a hierarchical approach evaluated against a baseline Yolov8 model. Detection accuracy metrics for isolated (“monop”), leaf nodes (“fenes”, “maryl”) and intermediate (“enrgg”) nodes are illustrated. Results are reported as hierarchical Precision hP, hierarchical Recall hR, and Fl -score Fl. The baseline approach can be a Yolov8 detection model, compared with successively adding a) pseudo-labels, b) hierarchical BCE loss, c) hierarchical routing.Table 1 - Precision and Recall Performance

[0066] These results demonstrate that an adaption to an object detection framework to hierarchical data can improve accuracy of downstream tasks such as key-point position accuracy. This improvement can be due to the fact that relaxing the loss from multi-class to hierarchical imposes the network to reason about the relationship between classes.

[0067] Turning now to FIG. 8, a flowchart of a method 700 for performing hierarchical object detection in surgical images is generally shown in accordance with one or more aspects. All or a portion of method 700 can be a computer-implemented method that is implemented, for example, by all or a portion of CAS system 100 of FIG. 1 and / or computer system 800 of FIG. 9.

[0068] At block 702, a surgical image 402 can be provided to a network 405 that includes a backbone 404 coupled to a bounding box head 406 and a classification head 410.

[0069] At block 704, a hierarchical binary cross-entropy loss 412 of an output of the classification head 410 is determined, where the output of the classification head 410 includes a multi-class probability vector with a hierarchical structure.

[0070] At block 706, the network 405 is trained using the hierarchical binary crossentropy loss 412 to reason with respect to relationships between classes of the hierarchical structure.

[0071] At block 708, class routing is performed at inference time to identify a label associated with a classification of the network 405 for an object in a bounding box defined by the bounding box head 406.

[0072] In some aspects, the surgical image 402 can be obtained from a video stream of a laparoscopic camera, the object can be a surgical instrument, and the hierarchical structure can be defined with respect to similar parts appearing on multiple instruments.

[0073] In some aspects, the network 405 can further include a pose head 408 coupled to the backbone 404, and the pose head 408 can provide a set of normalized key-point positions of the object.

[0074] In some aspects, the backbone 404 can include a convolutional neural network backbone with a feature pyramidal network.

[0075] In some aspects, the hierarchical binary cross-entropy loss 412 can be determined by analyzing a set of ancestors of a node of a hierarchy graph and adjusting ancestor and director successor values relative to a ground truth.

[0076] In some aspects, pseudo-labels can be used to transform a plurality of standard hierarchy-agnostic labels into a directed graph representing a hierarchy of classes. One or more objects can be re-categorized based on visibility and occlusion of the one or more objects in the surgical image 402.

[0077] In some aspects, the class routing can choose a node based on a prediction score of the node and ancestors of the node, an ambiguity determination between the node and one or more other nodes in a same level of hierarchy, and the node can be at a highest level of hierarchy when more than one node conforms.

[0078] In some aspects, a system includes a data store 320 including video data associated with a surgical procedure and a machine learning training system 325 for training a network 405 using surgical images 402 extracted from the video data. The system is configured to provide the surgical images 402 to the network 405 including a backbone 404 coupled to a bounding box head 406 and a classification head 410. The system is further configured to determine a hierarchical binary cross-entropy loss 412 of an output of the classification head 410, where the output of the classification head 410 includes a multi-class probability vector with a hierarchical structure. The system is also configured to train the network 405 using the hierarchical binary cross-entropy loss 412 to reason with respect to relationships between classes of the hierarchical structure.

[0079] In some aspects, the object can be a surgical instrument, and the hierarchical structure can be defined with respect to similar parts appearing on multiple instruments.

[0080] In some aspects, the network 405 can further include a pose head 408 coupled to the backbone 404, and the pose head 408 can provide a set of normalized key-point positions of the object.

[0081] According to an aspect, a computer program product is provided and includes a memory device having computer executable instructions stored thereon, which when executed by one or more processors cause the one or more processors to perform a plurality of operations for performing hierarchical object detection in surgical images 402. The plurality of operations include providing a surgical image 402 to a network 405 including a backbone 404 coupled to a classification head 410, determining a hierarchical binary cross-entropy loss 412 of an output of the classification head 410, where the output of the classification head 410 includes a multi-class probability vector with a hierarchical structure, training the network 405 using the hierarchical binary crossentropy loss 412 to reason with respect to relationships between classes of the hierarchical structure, and performing class routing at inference time to identify a label associated with a classification of the network 405 for an object.

[0082] The processing shown in FIG. 8 is not intended to indicate that the operations are to be executed in any particular order or that all of the operations shown in FIG. 8 are to be included in every case. Additionally, the processing shown in FIG. 8 can include any suitable number of additional operations.

[0083] Turning now to FIG. 9, a computer system 800 is generally shown in accordance with an aspect. The computer system 800 can be an electronic computer framework comprising and / or employing any number and combination of computing devices and networks utilizing various communication technologies, as described herein. The computer system 800 can be easily scalable, extensible, and modular, with the ability to change to different services or reconfigure some features independently of others. The computer system 800 may be, for example, a server, desktop computer, laptop computer, tablet computer, or smartphone. In some examples, computer system 800 may be a cloud computing node. Computer system 800 may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer system. Generally, program modules may include routines, programs, objects, components, logic, data structures, and so on that perform particular tasks or implementparticular abstract data types. Computer system 800 may be practiced in distributed cloud computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed cloud computing environment, program modules may be located in both local and remote computer system storage media, including memory storage devices.

[0084] As shown in FIG. 9, the computer system 800 has one or more central processing units (CPU(s)) 801a, 801b, 801c, etc. (collectively or generically referred to as processor(s) 801). The processors 801 can be a single-core processor, multi-core processor, computing cluster, or any number of other configurations. The processors 801 can be any type of circuitry capable of executing instructions. The processors 801, also referred to as processing circuits, are coupled via a system bus 802 to a system memory 803 and various other components. The system memory 803 can include one or more memory devices, such as read-only memory (ROM) 804 and a random-access memory (RAM) 805. The ROM 804 is coupled to the system bus 802 and may include a basic input / output system (BIOS), which controls certain basic functions of the computer system 800. The RAM is read-write memory coupled to the system bus 802 for use by the processors 801. The system memory 803 provides temporary memory space for operations of said instructions during operation. The system memory 803 can include random access memory (RAM), read-only memory, flash memory, or any other suitable memory systems.

[0085] The computer system 800 comprises an input / output (I / O) adapter 806 and a communications adapter 807 coupled to the system bus 802. The I / O adapter 806 may be a small computer system interface (SCSI) adapter that communicates with a hard disk 808 and / or any other similar component. The I / O adapter 806 and the hard disk 808 are collectively referred to herein as a mass storage 810.

[0086] Software 811 for execution on the computer system 800 may be stored in the mass storage 810. The mass storage 810 is an example of a tangible storage medium readable by the processors 801, where the software 811 is stored as instructions forexecution by the processors 801 to cause the computer system 800 to operate, such as is described hereinbelow with respect to the various Figures. Examples of computer program product and the execution of such instruction is discussed herein in more detail. The communications adapter 807 interconnects the system bus 802 with a network 812, which may be an outside network, enabling the computer system 800 to communicate with other such systems. In one aspect, a portion of the system memory 803 and the mass storage 810 collectively store an operating system, which may be any appropriate operating system to coordinate the functions of the various components shown in FIG. 9.

[0087] Additional input / output devices are shown as connected to the system bus 802 via a display adapter 815 and an interface adapter 816. In one aspect, the adapters 806, 807, 815, and 816 may be connected to one or more VO buses that are connected to the system bus 802 via an intermediate bus bridge (not shown). A display 819 (e.g., a screen or a display monitor) is connected to the system bus 802 by a display adapter 815, which may include a graphics controller to improve the performance of graphics-intensive applications and a video controller. A keyboard, a mouse, a touchscreen, one or more buttons, a speaker, etc., can be interconnected to the system bus 802 via the interface adapter 816, which may include, for example, a Super I / O chip integrating multiple device adapters into a single integrated circuit. Suitable VO buses for connecting peripheral devices such as hard disk controllers, network adapters, and graphics adapters typically include common protocols, such as the Peripheral Component Interconnect (PCI). Thus, as configured in FIG. 9, the computer system 800 includes processing capability in the form of the processors 801, and storage capability including the system memory 803 and the mass storage 810, input means such as the buttons, touchscreen, and output capability including the speaker 823 and the display 819.

[0088] In some aspects, the communications adapter 807 can transmit data using any suitable interface or protocol, such as the internet small computer system interface, among others. The network 812 may be a cellular network, a radio network, a wide area network (WAN), a local area network (LAN), or the Internet, among others. An externalcomputing device may connect to the computer system 800 through the network 812. In some examples, an external computing device may be an external web server or a cloud computing node.

[0089] It is to be understood that the block diagram of FIG. 9 is not intended to indicate that the computer system 800 is to include all of the components shown in FIG.9. Rather, the computer system 800 can include any appropriate fewer or additional components not illustrated in FIG. 9 (e.g., additional memory components, embedded controllers, modules, additional network interfaces, etc.). Further, the aspects described herein with respect to computer system 800 may be implemented with any appropriate logic, wherein the logic, as referred to herein, can include any suitable hardware (e.g., a processor, an embedded controller, or an application-specific integrated circuit, among others), software (e.g., an application, among others), firmware, or any suitable combination of hardware, software, and firmware, in various aspects. Various aspects can be combined to include two or more of the aspects described herein.

[0090] Aspects disclosed herein may be a system, a method, and / or a computer program product at any possible technical detail level of integration. The computer program product may include a computer-readable storage medium (or media) having computer-readable program instructions thereon for causing a processor to carry out various aspects.

[0091] The computer-readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. The computer-readable storage medium may be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non- exhaustive list of more specific examples of the computer-readable storage medium includes the following: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portablecompact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanically encoded device, such as punch-cards or raised structures in a groove having instructions recorded thereon, and any suitable combination of the foregoing. A computer-readable storage medium, as used herein, is not to be construed as being transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission media (e.g., light pulses passing through a fiber-optic cable), or electrical signals transmitted through a wire.

[0092] Computer-readable program instructions described herein can be downloaded to respective computing / processing devices from a computer-readable storage medium or to an external computer or external storage device via a network, for example, the Internet, a local area network, a wide area network, and / or a wireless network. The network may comprise copper transmission cables, optical transmission fibers, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives computer-readable program instructions from the network and forwards the computer- readable program instructions for storage in a computer-readable storage medium within the respective computing / processing device.

[0093] Computer-readable program instructions for carrying out operations of the present disclosure may be assembler instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state-setting data, configuration data for integrated circuitry, or either source-code or object code written in any combination of one or more programming languages, including an object-oriented programming language, such as Smalltalk, C++, high-level languages such as Python, or the like, and procedural programming languages, such as the “C” programming language or similar programming languages. The computer-readable program instructions may execute entirely on the user’s computer, partly on the user’s computer, as a stand-alone software package, partly on the user’scomputer and partly on a remote computer, or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user’s computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet Service Provider). In some aspects, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate arrays (FPGA), or programmable logic arrays (PLA) may execute the computer-readable program instruction by utilizing state information of the computer-readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present disclosure.

[0094] Aspects are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to aspects of the disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0095] These computer-readable program instructions may be provided to a processor of a computer system, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer-readable program instructions may also be stored in a computer-readable storage medium that can direct a computer, a programmable data processing apparatus, and / or other devices to function in a particular manner, such that the computer-readable storage medium having instructions stored therein comprises an article of manufacture including instructions which implement aspects of the function / act specified in the flowchart and / or block diagram block or blocks.

[0096] The computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer-implemented process, such that the instructions which execute on the computer, other programmable apparatus, or other device implement the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0097] The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various aspects. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the blocks may occur out of the order noted in the Figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and / or flowchart illustration, and combinations of blocks in the block diagrams and / or flowchart illustration, can be implemented by special purpose hardware-based systems that perform the specified functions or acts or carry out combinations of special purpose hardware and computer instructions.

[0098] The descriptions of the various aspects have been presented for purposes of illustration but are not intended to be exhaustive or limited to the aspects disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described aspects. The terminology used herein was chosen to best explain the principles of the aspects, the practical application or technical improvement over technologies found in the marketplace, or to enable others of ordinary skill in the art to understand the aspects described herein.

[0099] Various aspects are described herein with reference to the related drawings. Alternative aspects can be devised without departing from the scope of this disclosure. Various connections and positional relationships (e.g., over, below, adjacent, etc.) are set forth between elements in the following description and in the drawings. These connections and / or positional relationships, unless specified otherwise, can be direct or indirect, and the present disclosure is not intended to be limiting in this respect.Accordingly, a coupling of entities can refer to either a direct or an indirect coupling, and a positional relationship between entities can be a direct or indirect positional relationship. Moreover, the various tasks and process steps described herein can be incorporated into a more comprehensive procedure or process having additional steps or functionality not described in detail herein.

[0100] The following definitions and abbreviations are to be used for the interpretation of the claims and the specification. As used herein, the terms “comprises,” “comprising,” “includes,” “including,” “has,” “having,” “contains,” or “containing,” or any other variation thereof are intended to cover a non-exclusive inclusion. For example, a composition, a mixture, process, method, article, or apparatus that comprises a list of elements is not necessarily limited to only those elements but can include other elements not expressly listed or inherent to such composition, mixture, process, method, article, or apparatus.

[0101] Additionally, the term “exemplary” is used herein to mean “serving as an example, instance or illustration.” Any aspect or design described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other aspects or designs. The terms “at least one” and “one or more” may be understood to include any integer number greater than or equal to one, i.e., one, two, three, four, etc. The terms “a plurality” may be understood to include any integer number greater than or equal to two, i.e., two, three, four, five, etc. The term “connection” may include both an indirect “connection” and a direct “connection.”

[0102] The terms “about,” “substantially,” “approximately,” and variations thereof are intended to include the degree of error associated with measurement of the particular quantity based upon the equipment available at the time of filing the application. For example, “about” can include a range of ± 8% or 5%, or 2% of a given value.

[0103] For the sake of brevity, conventional techniques related to making and using aspects may or may not be described in detail herein. In particular, various aspects of computing systems and specific computer programs to implement the various technical features described herein are well known. Accordingly, in the interest of brevity, many conventional implementation details are only mentioned briefly herein or are omitted entirely without providing the well-known system and / or process details.

[0104] It should be understood that various aspects, and / or parts of the aspects, disclosed herein may be combined in different combinations than the combinations specifically presented in the description and accompanying drawings. It should also be understood that, depending on the example, certain acts or events of any of the processes or methods described herein may be performed in a different sequence, may be added, merged, or left out altogether (e.g., all described acts or events may not be necessary to carry out the techniques). In addition, while certain aspects of this disclosure are described as being performed by a single module or unit for purposes of clarity, it should be understood that the techniques of this disclosure may be performed by a combination of units or modules associated with, for example, a medical device.

[0105] In one or more examples, the described techniques may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functions may be stored as one or more instructions or code on a computer-readable medium and executed by a hardware-based processing unit. Computer-readable media may include non-transitory computer-readable media, which corresponds to a tangible medium, such as data storage media (e.g., RAM, ROM, EEPROM, flash memory, or any other medium that can be used to store desired program code in the form of instructions or data structures and that can be accessed by a computer).

[0106] Instructions may be executed by one or more processors, such as one or more digital signal processors (DSPs), graphics processing units (GPUs), microprocessors, application-specific integrated circuits (ASICs), field programmable logic arrays (FPGAs), or other equivalent integrated or discrete logic circuitry. Accordingly, the term “processor” as used herein may refer to any of the foregoing structure or any other physical structure suitable for implementation of the described techniques. Also, the techniques could be fully implemented in one or more circuits or logic elements.

[0107] While the invention has been described with reference to aspects, it should be understood by those skilled in the art that various changes may be made and equivalents may be substituted for elements thereof without departing from the scope of the invention. Moreover, the aspects or parts of the aspects may be combined in whole or in part without departing from the scope of the invention. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from the scope thereof. Therefore, it is intended that the invention not be limited to the particular aspects disclosed as contemplated for carrying out this invention, but that the invention will include all aspects falling within the scope of the appended claims. Moreover, unless specifically stated any use of the terms first, second, etc. do not denote any order or importance, but rather the terms first, second, etc. are used to distinguish one element from another.

Claims

CLAIMSWhat is claimed is:

1. A computer-implemented method for performing hierarchical object detection in surgical images, comprising: providing a surgical image to a network comprising a backbone coupled to a bounding box head and a classification head; determining a hierarchical binary cross-entropy loss of an output of the classification head, wherein the output of the classification head comprises a multi-class probability vector with a hierarchical structure; training the network using the hierarchical binary cross-entropy loss to reason with respect to relationships between classes of the hierarchical structure; and performing class routing at inference time to identify a label associated with a classification of the network for an object in a bounding box defined by the bounding box head.

2. The computer-implemented method of claim 1, wherein the surgical image is obtained from a video stream of a laparoscopic camera, the object is a surgical instrument, and the hierarchical structure is defined with respect to similar parts appearing on multiple instruments.

3. The computer-implemented method of claim 1 or claim 2, wherein the network further comprises a pose head coupled to the backbone, and the pose head provides a set of normalized key -point positions of the object.

4. The computer-implemented method of any one of claims 1 to 3, wherein the backbone comprises a convolutional neural network backbone with a feature pyramidal network.

5. The computer-implemented method of any preceding claim, wherein the hierarchical binary cross-entropy loss is determined by analyzing a set of ancestors of a node of a hierarchy graph and adjusting ancestor and director successor values relative to a ground truth.

6. The computer-implemented method of any preceding claim, further comprising: using pseudo-labels to transform a plurality of standard hierarchy-agnostic labels into a directed graph representing a hierarchy of classes; and re-categorizing one or more objects based on visibility and occlusion of the one or more objects in the surgical image.

7. The computer-implemented method of any preceding claim, wherein the class routing chooses a node based on a prediction score of the node and ancestors of the node, an ambiguity determination between the node and one or more other nodes in a same level of hierarchy, and the node is at a highest level of hierarchy when more than one node conforms.

8. A system comprising: a data store comprising video data associated with a surgical procedure; and a machine learning training system for training a network using surgical images extracted from the video data, the system configured to: provide the surgical images to the network comprising a backbone coupled to a bounding box head and a classification head; determine a hierarchical binary cross-entropy loss of an output of the classification head, wherein the output of the classification head comprises a multi-class probability vector with a hierarchical structure; and train the network using the hierarchical binary cross-entropy loss to reason with respect to relationships between classes of the hierarchical structure.

9. The system of claim 8, wherein the object is a surgical instrument, and the hierarchical structure is defined with respect to similar parts appearing on multiple instruments.

10. The system of claim 8 or claim 9, wherein the network further comprises a pose head coupled to the backbone, and the pose head provides a set of normalized key-point positions of the object.

11. The system of any of claims 8 to 10, wherein the backbone comprises a convolutional neural network backbone with a feature pyramidal network.

12. The system of any of claims 8 to 11, wherein the hierarchical binary cross-entropy loss is determined by analyzing a set of ancestors of a node of a hierarchy graph and adjusting ancestor and director successor values relative to a ground truth.

13. The system of any of claims 8 to 12, wherein pseudo-labels are used to transform a plurality of standard hierarchy-agnostic labels into a directed graph representing a hierarchy of classes.

14. The system of claim 13, wherein one or more objects are re-categorized based on visibility and occlusion of the one or more objects in the surgical images.

15. A computer program product comprising a memory device having computer executable instructions stored thereon, which when executed by one or more processors cause the one or more processors to perform a plurality of operations for performing hierarchical object detection in surgical images, the plurality of operations comprising: providing a surgical image to a network comprising a backbone coupled to a classification head; determining a hierarchical binary cross-entropy loss of an output of theclassification head, wherein the output of the classification head comprises a multi-class probability vector with a hierarchical structure; training the network using the hierarchical binary cross-entropy loss to reason with respect to relationships between classes of the hierarchical structure; and performing class routing at inference time to identify a label associated with a classification of the network for an object.

16. The computer program product of claim 15, wherein the object is a surgical instrument, and the hierarchical structure is defined with respect to similar parts appearing on multiple instruments.

17. The computer program product of claim 15 or claim 16, wherein the network further comprises a pose head coupled to the backbone, and the pose head provides a set of normalized key -point positions of the object.

18. The computer program product of any of claims 15 to 17, wherein the backbone comprises a convolutional neural network backbone with a feature pyramidal network.

19. The computer program product of any of claims 15 to 18, wherein the hierarchical binary cross-entropy loss is determined by analyzing a set of ancestors of a node of a hierarchy graph and adjusting ancestor and director successor values relative to a ground truth.

20. The computer program product of any of claims 15 to 19, wherein the class routing chooses a node based on a prediction score of the node and ancestors of the node, an ambiguity determination between the node and one or more other nodes in a same level of hierarchy, and the node is at a highest level of hierarchy when more than one node conforms.