Unsupervised Trajectory Segmentation for Surgical Gesture Recognition in Robotic Training

Artificial intelligence integration in surgery through hand and instrument tracking: a systematic literature review

Objective

This systematic literature review of the integration of artificial intelligence (AI) applications in surgical practice through hand and instrument tracking provides an overview of recent advancements and analyzes current literature on the intersection of surgery with AI. Distinct AI algorithms and specific applications in surgical practice are also examined.

Methods

An advanced search using medical subject heading terms was conducted in Medline (via PubMed), SCOPUS, and Embase databases for articles published in English. A strict selection process was performed, adhering to PRISMA guidelines.

Results

A total of 225 articles were retrieved. After screening, 77 met inclusion criteria and were included in the review. Use of AI algorithms in surgical practice was uncommon during 2013-2017 but has gained significant popularity since 2018. Deep learning algorithms (n = 62) are increasingly preferred over traditional machine learning algorithms (n = 15). These technologies are used in surgical fields such as general surgery (n = 19), neurosurgery (n = 10), and ophthalmology (n = 9). The most common functional sensors and systems used were prerecorded videos (n = 29), cameras (n = 21), and image datasets (n = 7). The most common applications included laparoscopic (n = 13), robotic-assisted (n = 13), basic (n = 12), and endoscopic (n = 8) surgical skills training, as well as surgical simulation training (n = 8).

Conclusion

AI technologies can be tailored to address distinct needs in surgical education and patient care. The use of AI in hand and instrument tracking improves surgical outcomes by optimizing surgical skills training. It is essential to acknowledge the current technical and social limitations of AI and work toward filling those gaps in future studies.

Segment, Compare, and Learn: Creating Movement Libraries of Complex Task for Learning from Demonstration

Motion primitives are a highly useful and widely employed tool in the field of Learning from Demonstration (LfD). However, obtaining a large number of motion primitives can be a tedious process, as they typically need to be generated individually for each task to be learned. To address this challenge, this work presents an algorithm for acquiring robotic skills through automatic and unsupervised segmentation. The algorithm divides tasks into simpler subtasks and generates motion primitive libraries that group common subtasks for use in subsequent learning processes. Our algorithm is based on an initial segmentation step using a heuristic method, followed by probabilistic clustering with Gaussian Mixture Models. Once the segments are obtained, they are grouped using Gaussian Optimal Transport on the Gaussian Processes (GPs) of each segment group, comparing their similarities through the energy cost of transforming one GP into another. This process requires no prior knowledge, it is entirely autonomous, and supports multimodal information. The algorithm enables generating trajectories suitable for robotic tasks, establishing simple primitives that encapsulate the structure of the movements to be performed. Its effectiveness has been validated in manipulation tasks with a real robot, as well as through comparisons with state-of-the-art algorithms.

Systematic review of machine learning applications using nonoptical motion tracking in surgery

This systematic review explores machine learning (ML) applications in surgical motion analysis using non-optical motion tracking systems (NOMTS), alone or with optical methods. It investigates objectives, experimental designs, model effectiveness, and future research directions. From 3632 records, 84 studies were included, with Artificial Neural Networks (38%) and Support Vector Machines (11%) being the most common ML models. Skill assessment was the primary objective (38%). NOMTS used included internal device kinematics (56%), electromagnetic (17%), inertial (15%), mechanical (11%), and electromyography (1%) sensors. Surgical settings were robotic (60%), laparoscopic (18%), open (16%), and others (6%). Procedures focused on bench-top tasks (67%), clinical models (17%), clinical simulations (9%), and non-clinical simulations (7%). Over 90% accuracy was achieved in 36% of studies. Literature shows NOMTS and ML can enhance surgical precision, assessment, and training. Future research should advance ML in surgical environments, ensure model interpretability and reproducibility, and use larger datasets for accurate evaluation.

Wi-AM: Enabling Cross-Domain Gesture Recognition with Commodity Wi-Fi

RF-based gesture recognition systems outperform computer vision-based systems in terms of user privacy. The integration of Wi-Fi sensing and deep learning has opened new application areas for intelligent multimedia technology. Although promising, existing systems have multiple limitations: (1) they only work well in a fixed domain; (2) when working in a new domain, they require the recollection of a large amount of data. These limitations either lead to a subpar cross-domain performance or require a huge amount of human effort, impeding their widespread adoption in practical scenarios. We propose Wi-AM, a privacy-preserving gesture recognition framework, to address the above limitations. Wi-AM can accurately recognize gestures in a new domain with only one sample. To remove irrelevant disturbances induced by interfering domain factors, we design a multi-domain adversarial scheme to reduce the differences in data distribution between different domains and extract the maximum amount of transferable features related to gestures. Moreover, to quickly adapt to an unseen domain with only a few samples, Wi-AM adopts a meta-learning framework to fine-tune the trained model into a new domain with a one-sample-per-gesture manner while achieving an accurate cross-domain performance. Extensive experiments in a real-world dataset demonstrate that Wi-AM can recognize gestures in an unseen domain with average accuracy of 82.13% and 86.76% for 1 and 3 data samples.

PEg TRAnsfer Workflow recognition challenge report: Do multimodal data improve recognition?

BACKGROUND AND OBJECTIVE

In order to be context-aware, computer-assisted surgical systems require accurate, real-time automatic surgical workflow recognition. In the past several years, surgical video has been the most commonly-used modality for surgical workflow recognition. But with the democratization of robot-assisted surgery, new modalities, such as kinematics, are now accessible. Some previous methods use these new modalities as input for their models, but their added value has rarely been studied. This paper presents the design and results of the "PEg TRAnsfer Workflow recognition" (PETRAW) challenge with the objective of developing surgical workflow recognition methods based on one or more modalities and studying their added value.

METHODS

The PETRAW challenge included a data set of 150 peg transfer sequences performed on a virtual simulator. This data set included videos, kinematic data, semantic segmentation data, and annotations, which described the workflow at three levels of granularity: phase, step, and activity. Five tasks were proposed to the participants: three were related to the recognition at all granularities simultaneously using a single modality, and two addressed the recognition using multiple modalities. The mean application-dependent balanced accuracy (AD-Accuracy) was used as an evaluation metric to take into account class balance and is more clinically relevant than a frame-by-frame score.

RESULTS

Seven teams participated in at least one task with four participating in every task. The best results were obtained by combining video and kinematic data (AD-Accuracy of between 93% and 90% for the four teams that participated in all tasks).

CONCLUSION

The improvement of surgical workflow recognition methods using multiple modalities compared with unimodal methods was significant for all teams. However, the longer execution time required for video/kinematic-based methods(compared to only kinematic-based methods) must be considered. Indeed, one must ask if it is wise to increase computing time by 2000 to 20,000% only to increase accuracy by 3%. The PETRAW data set is publicly available at www.synapse.org/PETRAW to encourage further research in surgical workflow recognition.

Computer-assisted approaches for measuring, segmenting, and analyzing functional upper extremity movement: a narrative review of the current state, limitations, and future directions

The analysis of functional upper extremity (UE) movement kinematics has implications across domains such as rehabilitation and evaluating job-related skills. Using movement kinematics to quantify movement quality and skill is a promising area of research but is currently not being used widely due to issues associated with cost and the need for further methodological validation. Recent developments by computationally-oriented research communities have resulted in potentially useful methods for evaluating UE function that may make kinematic analyses easier to perform, generally more accessible, and provide more objective information about movement quality, the importance of which has been highlighted during the COVID-19 pandemic. This narrative review provides an interdisciplinary perspective on the current state of computer-assisted methods for analyzing UE kinematics with a specific focus on how to make kinematic analyses more accessible to domain experts. We find that a variety of methods exist to more easily measure and segment functional UE movement, with a subset of those methods being validated for specific applications. Future directions include developing more robust methods for measurement and segmentation, validating these methods in conjunction with proposed kinematic outcome measures, and studying how to integrate kinematic analyses into domain expert workflows in a way that improves outcomes.

Effects of automated skill assessment on robotic surgery training

BACKGROUND

Several automated skill-assessment approaches have been proposed for robotic surgery, but their utility is not well understood. This article investigates the effects of one machine-learning-based skill-assessment approach on psychomotor skill development in robotic surgery training.

METHODS

N = 29 trainees (medical students and residents) with no robotic surgery experience performed five trials of inanimate peg transfer with an Intuitive Surgical da Vinci Standard robot. Half of the participants received no post-trial feedback. The other half received automatically calculated scores from five Global Evaluative Assessment of Robotic Skill domains post-trial.

RESULTS

There were no significant differences between the groups regarding overall improvement or skill improvement rate. However, participants who received post-trial feedback rated their overall performance improvement significantly lower than participants who did not receive feedback.

CONCLUSIONS

These findings indicate that automated skill evaluation systems might improve trainee self-awareness but not accelerate early stage psychomotor skill development in robotic surgery training.

A virtual surgical prototype system based on gesture recognition for virtual surgical training in maxillofacial surgery

Background

Virtual reality (VR) technology is an ideal alternative of operation training and surgical teaching. However, virtual surgery is usually carried out using the mouse or data gloves, which affects the authenticity of virtual operation. A virtual surgery system with gesture recognition and real-time image feedback was explored to realize more authentic immersion.

Method

Gesture recognition technology proposed with an efficient and real-time algorithm and high fidelity was explored. The recognition of hand contour, palm and fingertip was firstly realized by hand data extraction. Then, an Support Vector Machine classifier was utilized to classify and recognize common gestures after extraction of feature recognition. The algorithm of collision detection adopted Axis Aligned Bounding Box binary tree to build hand and scalpel collision models. What’s more, nominal radius theorem (NRT) and separating axis theorem (SAT) were applied for speeding up collision detection. Based on the maxillofacial virtual surgical system we proposed before, the feasibility of integration of the above technologies in this prototype system was evaluated.

Results

Ten kinds of signal static gestures were designed to test gesture recognition algorithms. The accuracy of gestures recognition is more than 80%, some of which were over 90%. The generation speed of collision detection model met the software requirements with the method of NRT and SAT. The response time of gesture] recognition was less than 40 ms, namely the speed of hand gesture recognition system was greater than 25 Hz. On the condition of integration of hand gesture recognition, typical virtual surgical procedures including grabbing a scalpel, puncture site selection, virtual puncture operation and incision were carried out with realization of real-time image feedback.

Conclusion

Based on the previous maxillofacial virtual surgical system that consisted of VR, triangular mesh collision detection and maxillofacial biomechanical model construction, the integration of hand gesture recognition was a feasible method to improve the interactivity and immersion of virtual surgical operation training.

Digital Media Design for Dynamic Gesture Interaction with Image Processing

With the rapid development of human-computer interaction technology and research in the fields of ergonomics and user experience, people have increasingly demanded robot availability and ease of use. The rapid development of modern computer computing and digital media technology has made people’s contact and interaction with computers more and more frequent, and they have appeared more in people’s daily lives. Therefore, the convenience and freedom of computer communication have been proposed. Based on the above background, the research content of this paper is the design of dynamic gesture interactive digital media based on image processing. This paper proposes a design scheme of an interactive touch system based on image processing and a dynamic gesture recognition method and uses the time difference method to experimentally simulate the system proposed in this paper. Image processing is the technique of analyzing images with a computer to achieve the desired results. Image processing technology generally includes three parts: image compression, enhancement and restoration, and matching, description, and recognition. The experimental results show that, for the same motion trajectory, the accuracy depends on the complexity of the gesture degree. And the system’s recognition accuracy rate is always maintained above 98%, confirming that the system has the performance requirements of high recognition accuracy, fast response speed, and stable work. Finally, the system was tested for performance and function, which verified that the system meets the real-time requirements. For a particular gesture operator, the accuracy rate has a great relationship with the standard degree of its operation. If the similarity with the model is high, you can achieve a high recognition rate.

A Multilayer and Multimodal-Fusion Architecture for Simultaneous Recognition of Endovascular Manipulations and Assessment of Technical Skills

The clinical success of the percutaneous coronary intervention (PCI) is highly dependent on endovascular manipulation skills and dexterous manipulation strategies of interventionalists. However, the analysis of endovascular manipulations and related discussion for technical skill assessment are limited. In this study, a multilayer and multimodal-fusion architecture is proposed to recognize six typical endovascular manipulations. The synchronously acquired multimodal motion signals from ten subjects are used as the inputs of the architecture independently. Six classification-based and two rule-based fusion algorithms are evaluated for performance comparisons. The recognition metrics under the determined architecture are further used to assess technical skills. The experimental results indicate that the proposed architecture can achieve the overall accuracy of 96.41%, much higher than that of a single-layer recognition architecture (92.85%). In addition, the multimodal fusion brings significant performance improvement in comparison with single-modal schemes. Furthermore, the K -means-based skill assessment can obtain an accuracy of 95% to cluster the attempts made by different skill-level groups. These hopeful results indicate the great possibility of the architecture to facilitate clinical skill assessment and skill learning.

Preliminary design and evaluation of a remote tele-mentoring system for minimally invasive surgery

BACKGROUND

Tele-mentoring during surgery facilitates the transfer of surgical knowledge from a mentor (specialist surgeon) to a mentee (operating surgeon). The aim of this work is to develop a tele-mentoring system tailored for minimally invasive surgery (MIS) where the mentor can remotely demonstrate to the mentee the required motion of the surgical instruments.

METHODS

A remote tele-mentoring system is implemented that generates visual cues in the form of virtual surgical instrument motion overlaid onto the live view of the operative field. The technical performance of the system is evaluated in a simulated environment, where the operating room and the central location of the mentor were physically located in different countries and connected over the internet. In addition, a user study was performed to assess the system as a mentoring tool.

RESULTS

On average, it took 260 ms to send a view of the operative field of 1920 × 1080 resolution from the operating room to the central location of the mentor and an average of 132 ms to receive the motion of virtual surgical instruments from the central location to the operating room. The user study showed that it is feasible for the mentor to demonstrate and for the mentee to understand and replicate the motion of surgical instruments.

CONCLUSION

The work demonstrates the feasibility of transferring information over the internet from a mentor to a mentee in the form of virtual surgical instruments. Their motion is overlaid onto the live view of the operative field enabling real-time interactions between both the surgeons.

Artificial Intelligence and Surgical Education: A Systematic Scoping Review of Interventions

OBJECTIVE

To synthesize peer-reviewed evidence related to the use of artificial intelligence (AI) in surgical education DESIGN: We conducted and reported a scoping review according to the standards outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analysis with extension for Scoping Reviews guideline and the fourth edition of the Joanna Briggs Institute Reviewer's Manual. We systematically searched eight interdisciplinary databases including MEDLINE-Ovid, ERIC, EMBASE, CINAHL, Web of Science: Core Collection, Compendex, Scopus, and IEEE Xplore. Databases were searched from inception until the date of search on April 13, 2021.

SETTING/PARTICIPANTS

We only examined original, peer-reviewed interventional studies that self-described as AI interventions, focused on medical education, and were relevant to surgical trainees (defined as medical or dental students, postgraduate residents, or surgical fellows) within the title and abstract (see Table 2). Animal, cadaveric, and in vivo studies were not eligible for inclusion.

RESULTS

After systematically searching eight databases and 4255 citations, our scoping review identified 49 studies relevant to artificial intelligence in surgical education. We found diverse interventions related to the evaluation of surgical competency, personalization of surgical education, and improvement of surgical education materials across surgical specialties. Many studies used existing surgical education materials, such as the Objective Structured Assessment of Technical Skills framework or the JHU-ISI Gesture and Skill Assessment Working Set database. Though most studies did not provide outcomes related to the implementation in medical schools (such as cost-effective analyses or trainee feedback), there are numerous promising interventions. In particular, many studies noted high accuracy in the objective characterization of surgical skill sets. These interventions could be further used to identify at-risk surgical trainees or evaluate teaching methods.

CONCLUSIONS

There are promising applications for AI in surgical education, particularly for the assessment of surgical competencies, though further evidence is needed regarding implementation and applicability.

Review of automated performance metrics to assess surgical technical skills in robot-assisted laparoscopy

INTRODUCTION

Robot-assisted laparoscopy is a safe surgical approach with several studies suggesting correlations between complication rates and the surgeon's technical skills. Surgical skills are usually assessed by questionnaires completed by an expert observer. With the advent of surgical robots, automated surgical performance metrics (APMs)-objective measures related to instrument movements-can be computed. The aim of this systematic review was thus to assess APMs use in robot-assisted laparoscopic procedures. The primary outcome was the assessment of surgical skills by APMs and the secondary outcomes were the association between APM and surgeon parameters and the prediction of clinical outcomes.

METHODS

A systematic review following the PRISMA guidelines was conducted. PubMed and Scopus electronic databases were screened with the query "robot-assisted surgery OR robotic surgery AND performance metrics" between January 2010 and January 2021. The quality of the studies was assessed by the medical education research study quality instrument. The study settings, metrics, and applications were analysed.

RESULTS

The initial search yielded 341 citations of which 16 studies were finally included. The study settings were either simulated virtual reality (VR) (4 studies) or real clinical environment (12 studies). Data to compute APMs were kinematics (motion tracking), and system and specific events data (actions from the robot console). APMs were used to differentiate expertise levels, and thus validate VR modules, predict outcomes, and integrate datasets for automatic recognition models. APMs were correlated with clinical outcomes for some studies.

CONCLUSIONS

APMs constitute an objective approach for assessing technical skills. Evidence of associations between APMs and clinical outcomes remain to be confirmed by further studies, particularly, for non-urological procedures. Concurrent validation is also required.

MIcro-surgical anastomose workflow recognition challenge report

BACKGROUND AND OBJECTIVE

Automatic surgical workflow recognition is an essential step in developing context-aware computer-assisted surgical systems. Video recordings of surgeries are becoming widely accessible, as the operational field view is captured during laparoscopic surgeries. Head and ceiling mounted cameras are also increasingly being used to record videos in open surgeries. This makes videos a common choice in surgical workflow recognition. Additional modalities, such as kinematic data captured during robot-assisted surgeries, could also improve workflow recognition. This paper presents the design and results of the MIcro-Surgical Anastomose Workflow recognition on training sessions (MISAW) challenge whose objective was to develop workflow recognition models based on kinematic data and/or videos.

METHODS

The MISAW challenge provided a data set of 27 sequences of micro-surgical anastomosis on artificial blood vessels. This data set was composed of videos, kinematics, and workflow annotations. The latter described the sequences at three different granularity levels: phase, step, and activity. Four tasks were proposed to the participants: three of them were related to the recognition of surgical workflow at three different granularity levels, while the last one addressed the recognition of all granularity levels in the same model. We used the average application-dependent balanced accuracy (AD-Accuracy) as the evaluation metric. This takes unbalanced classes into account and it is more clinically relevant than a frame-by-frame score.

RESULTS

Six teams participated in at least one task. All models employed deep learning models, such as convolutional neural networks (CNN), recurrent neural networks (RNN), or a combination of both. The best models achieved accuracy above 95%, 80%, 60%, and 75% respectively for recognition of phases, steps, activities, and multi-granularity. The RNN-based models outperformed the CNN-based ones as well as the dedicated modality models compared to the multi-granularity except for activity recognition.

CONCLUSION

For high levels of granularity, the best models had a recognition rate that may be sufficient for applications such as prediction of remaining surgical time. However, for activities, the recognition rate was still low for applications that can be employed clinically. The MISAW data set is publicly available at http://www.synapse.org/MISAW to encourage further research in surgical workflow recognition.

Gesture Recognition in Robotic Surgery: A Review

OBJECTIVE

Surgical activity recognition is a fundamental step in computer-assisted interventions. This paper reviews the state-of-the-art in methods for automatic recognition of fine-grained gestures in robotic surgery focusing on recent data-driven approaches and outlines the open questions and future research directions.

METHODS

An article search was performed on 5 bibliographic databases with the following search terms: robotic, robot-assisted, JIGSAWS, surgery, surgical, gesture, fine-grained, surgeme, action, trajectory, segmentation, recognition, parsing. Selected articles were classified based on the level of supervision required for training and divided into different groups representing major frameworks for time series analysis and data modelling.

RESULTS

A total of 52 articles were reviewed. The research field is showing rapid expansion, with the majority of articles published in the last 4 years. Deep-learning-based temporal models with discriminative feature extraction and multi-modal data integration have demonstrated promising results on small surgical datasets. Currently, unsupervised methods perform significantly less well than the supervised approaches.

CONCLUSION

The development of large and diverse open-source datasets of annotated demonstrations is essential for development and validation of robust solutions for surgical gesture recognition. While new strategies for discriminative feature extraction and knowledge transfer, or unsupervised and semi-supervised approaches, can mitigate the need for data and labels, they have not yet been demonstrated to achieve comparable performance. Important future research directions include detection and forecast of gesture-specific errors and anomalies.

SIGNIFICANCE

This paper is a comprehensive and structured analysis of surgical gesture recognition methods aiming to summarize the status of this rapidly evolving field.

Surgical Hand Gesture Recognition Utilizing Electroencephalogram as Input to the Machine Learning and Network Neuroscience Algorithms

Surgical gestures detection can provide targeted, automated surgical skill assessment and feedback during surgical training for robot-assisted surgery (RAS). Several sources including surgical videos, robot tool kinematics, and an electromyogram (EMG) have been proposed to reach this goal. We aimed to extract features from electroencephalogram (EEG) data and use them in machine learning algorithms to classify robot-assisted surgical gestures. EEG was collected from five RAS surgeons with varying experience while performing 34 robot-assisted radical prostatectomies over the course of three years. Eight dominant hand and six non-dominant hand gesture types were extracted and synchronized with associated EEG data. Network neuroscience algorithms were utilized to extract functional brain network and power spectral density features. Sixty extracted features were used as input to machine learning algorithms to classify gesture types. The analysis of variance (ANOVA) F-value statistical method was used for feature selection and 10-fold cross-validation was used to validate the proposed method. The proposed feature set used in the extra trees (ET) algorithm classified eight gesture types performed by the dominant hand of five RAS surgeons with an accuracy of 90%, precision: 90%, sensitivity: 88%, and also classified six gesture types performed by the non-dominant hand with an accuracy of 93%, precision: 94%, sensitivity: 94%.

Automated Vision-Based Microsurgical Skill Analysis in Neurosurgery Using Deep Learning: Development and Preclinical Validation

BACKGROUND/OBJECTIVE

Technical skill acquisition is an essential component of neurosurgical training. Educational theory suggests that optimal learning and improvement in performance depends on the provision of objective feedback. Therefore, the aim of this study was to develop a vision-based framework based on a novel representation of surgical tool motion and interactions capable of automated and objective assessment of microsurgical skill.

METHODS

Videos were obtained from 1 expert, 6 intermediate, and 12 novice surgeons performing arachnoid dissection in a validated clinical model using a standard operating microscope. A mask region convolutional neural network framework was used to segment the tools present within the operative field in a recorded video frame. Tool motion analysis was achieved using novel triangulation metrics. Performance of the framework in classifying skill levels was evaluated using the area under the curve and accuracy. Objective measures of classifying the surgeons' skill level were also compared using the Mann-Whitney U test, and a value of P < 0.05 was considered statistically significant.

RESULTS

The area under the curve was 0.977 and the accuracy was 84.21%. A number of differences were found, which included experts having a lower median dissector velocity (P = 0.0004; 190.38 ms^-1 vs. 116.38 ms^-1), and a smaller inter-tool tip distance (median 46.78 vs. 75.92; P = 0.0002) compared with novices.

CONCLUSIONS

Automated and objective analysis of microsurgery is feasible using a mask region convolutional neural network, and a novel tool motion and interaction representation. This may support technical skills training and assessment in neurosurgery.

Computer Vision in the Surgical Operating Room

Background

Multiple types of surgical cameras are used in modern surgical practice and provide a rich visual signal that is used by surgeons to visualize the clinical site and make clinical decisions. This signal can also be used by artificial intelligence (AI) methods to provide support in identifying instruments, structures, or activities both in real-time during procedures and postoperatively for analytics and understanding of surgical processes.

Summary

In this paper, we provide a succinct perspective on the use of AI and especially computer vision to power solutions for the surgical operating room (OR). The synergy between data availability and technical advances in computational power and AI methodology has led to rapid developments in the field and promising advances.

Key Messages

With the increasing availability of surgical video sources and the convergence of technologies around video storage, processing, and understanding, we believe clinical solutions and products leveraging vision are going to become an important component of modern surgical capabilities. However, both technical and clinical challenges remain to be overcome to efficiently make use of vision-based approaches into the clinic.

Characterizing Functional Upper Extremity Movement in Haptic Virtual Environments

Haptic virtual environments have been used to assess cognitive and fine motor function. For tasks performed in physical environments, upper extremity movement is usually separated into reaching and object manipulation phases using fixed velocity thresholds. However, these thresholds can result in premature segmentation due to additional trajectory adjustments common in virtual environments. In this work, we address the issues of premature segmentation and the lack of a measure to characterize the spatial distribution of a trajectory while targeting an object. We propose a combined relative distance and velocity segmentation procedure and use principal component analysis (PCA) to capture the spatial distribution of the participant's targeting phase. Synthetic data and 3D motion data from twenty healthy adults were used to evaluate the methods with positive results. We found that these methods quantify motor skill improvement of healthy participants performing repeated trials of a haptic virtual environment task.

Quantifying surgeon maneuevers across experience levels through marker-less hand motion kinematics of simulated surgical tasks

This paper compares clinician hand motion for common suturing tasks across a range of experience levels and tissue types. Medical students (32), residents (41), attending surgeons (10), and retirees (2) were recorded on digital video while suturing on one of: foam, pig feet, or porcine bowel. Depending on time in position, each medical student, resident, and attending participant was classified as junior or senior, yielding six experience categories. This work focuses on trends associated with increasing tenure observed from those medical students (10), residents (15), and attendings (10) who sutured on foam, and draws comparison across tissue types where pertinent. Utilizing custom software, the two-dimensional location of each of the participant's hands were automatically recorded in every video frame, producing a rich spatiotemporal feature set. While suturing on foam, increasing clinician experience was associated with conserved path length per cycle of the non-dominant hand, significantly reducing from junior medical students (mean = 73.63 cm, sd = 33.21 cm) to senior residents (mean = 46.16 cm, sd = 14.03 cm, p = 0.015), and again between senior residents and senior attendings (mean = 30.84 cm, sd = 14.51 cm, p = 0.045). Despite similar maneuver rates, attendings also accelerated less with their non-dominant hand (mean = 16.27 cm/s², sd = 81.12 cm/s², p = 0.002) than senior residents (mean = 24.84 cm/s², sd = 68.29 cm/s², p = 0.002). While tying, medical students moved their dominant hands slower (mean = 4.39 cm/s, sd = 1.73 cm/s, p = 0.033) than senior residents (mean = 6.53 cm/s, sd = 2.52 cm/s). These results suggest that increased psychomotor performance during early training manifest through faster dominant hand function, while later increases are characterized by conserving energy and efficiently distributing work between hands. Incorporating this scalable video-based motion analysis into regular formative assessment routines may enable greater quality and consistency of feedback throughout a surgical career.

Offline identification of surgical deviations in laparoscopic rectopexy

OBJECTIVE

According to a meta-analysis of 7 studies, the median number of patients with at least one adverse event during the surgery is 14.4%, and a third of those adverse events were preventable. The occurrence of adverse events forces surgeons to implement corrective strategies and, thus, deviate from the standard surgical process. Therefore, it is clear that the automatic identification of adverse events is a major challenge for patient safety. In this paper, we have proposed a method enabling us to identify such deviations. We have focused on identifying surgeons' deviations from standard surgical processes due to surgical events rather than anatomic specificities. This is particularly challenging, given the high variability in typical surgical procedure workflows.

METHODS

We have introduced a new approach designed to automatically detect and distinguish surgical process deviations based on multi-dimensional non-linear temporal scaling with a hidden semi-Markov model using manual annotation of surgical processes. The approach was then evaluated using cross-validation.

RESULTS

The best results have over 90% accuracy. Recall and precision for event deviations, i.e. related to adverse events, are respectively below 80% and 40%. To understand these results, we have provided a detailed analysis of the incorrectly-detected observations.

CONCLUSION

Multi-dimensional non-linear temporal scaling with a hidden semi-Markov model provides promising results for detecting deviations. Our error analysis of the incorrectly-detected observations offers different leads in order to further improve our method.

SIGNIFICANCE

Our method demonstrated the feasibility of automatically detecting surgical deviations that could be implemented for both skill analysis and developing situation awareness-based computer-assisted surgical systems.

Robot complex motion learning based on unsupervised trajectory segmentation and movement primitives

This paper presents a robot skill acquisition framework for learning and reproducing humanoid trajectories with complex forms. A new unsupervised segmentation method is proposed to detect motion units in the demonstrated kinematic data using the concept of key points. To find the consistent features of trajectories, a Hidden Semi-Markov Model (HSMM) is used to identify key points common to all the demonstrations. Generalizing the motion units is achieved via a Probability-based Movement Primitive (PbMP), which encapsulates multiple trajectories into one model. Such a framework can generate trajectories suitable for robot execution with arbitrary shape and complexity from a small number of demonstrations, which greatly expands the application scenarios of robot programming by demonstration. The automatic segmentation process does not rely on a priori knowledge or models for specific tasks, and the generalized trajectory retains more consistent features than those produced by other algorithms. We demonstrate the effectiveness of the proposed framework through simulations and experiments.

Electrophysiological indicators of gesture perception

Electroencephalography (EEG) activity in the mu frequency band (8-13 Hz) is suppressed during both gesture performance and observation. However, it is not clear if or how particular characteristics within the kinematic execution of gestures map onto dynamic changes in mu activity. Mapping the time course of gesture kinematics onto that of mu activity could help understand which aspects of gestures capture attention and aid in the classification of communicative intent. In this work, we test whether the timing of inflection points within gesture kinematics predicts the occurrence of oscillatory mu activity during passive gesture observation. The timing for salient features of performed gestures in video stimuli was determined by isolating inflection points in the hands' motion trajectories. Participants passively viewed the gesture videos while continuous EEG data was collected. We used wavelet analysis to extract mu oscillations at 11 Hz and at central electrodes and occipital electrodes. We used linear regression to test for associations between the timing of inflection points in motion trajectories and mu oscillations that generalized across gesture stimuli. Separately, we also tested whether inflection point occurrences evoked mu/alpha responses that generalized across participants. Across all gestures and inflection points, and pooled across participants, peaks in 11 Hz EEG waveforms were detected 465 and 535 ms after inflection points at occipital and central electrodes, respectively. A regression model showed that inflection points in the motion trajectories strongly predicted subsequent mu oscillations ([Formula: see text]<0.01); effects were weaker and non-significant for low (17 Hz) and high (21 Hz) beta activity. When segmented by inflection point occurrence rather than stimulus onset and testing participants as a random effect, inflection points evoked mu and beta activity from 308 to 364 ms at central electrodes, and broad activity from 226 to 800 ms at occipital electrodes. The results suggest that inflection points in gesture trajectories elicit coordinated activity in the visual and motor cortices, with prominent activity in the mu/alpha frequency band and extending into the beta frequency band. The time course of activity indicates that visual processing drives subsequent activity in the motor cortex during gesture processing, with a lag of approximately 80 ms.

CAI4CAI: The Rise of Contextual Artificial Intelligence in Computer Assisted Interventions

Data-driven computational approaches have evolved to enable extraction of information from medical images with a reliability, accuracy and speed which is already transforming their interpretation and exploitation in clinical practice. While similar benefits are longed for in the field of interventional imaging, this ambition is challenged by a much higher heterogeneity. Clinical workflows within interventional suites and operating theatres are extremely complex and typically rely on poorly integrated intra-operative devices, sensors, and support infrastructures. Taking stock of some of the most exciting developments in machine learning and artificial intelligence for computer assisted interventions, we highlight the crucial need to take context and human factors into account in order to address these challenges. Contextual artificial intelligence for computer assisted intervention, or CAI4CAI, arises as an emerging opportunity feeding into the broader field of surgical data science. Central challenges being addressed in CAI4CAI include how to integrate the ensemble of prior knowledge and instantaneous sensory information from experts, sensors and actuators; how to create and communicate a faithful and actionable shared representation of the surgery among a mixed human-AI actor team; how to design interventional systems and associated cognitive shared control schemes for online uncertainty-aware collaborative decision making ultimately producing more precise and reliable interventions.

Real-time automatic surgical phase recognition in laparoscopic sigmoidectomy using the convolutional neural network-based deep learning approach

BACKGROUND

Automatic surgical workflow recognition is a key component for developing the context-aware computer-assisted surgery (CA-CAS) systems. However, automatic surgical phase recognition focused on colorectal surgery has not been reported. We aimed to develop a deep learning model for automatic surgical phase recognition based on laparoscopic sigmoidectomy (Lap-S) videos, which could be used for real-time phase recognition, and to clarify the accuracies of the automatic surgical phase and action recognitions using visual information.

METHODS

The dataset used contained 71 cases of Lap-S. The video data were divided into frame units every 1/30 s as static images. Every Lap-S video was manually divided into 11 surgical phases (Phases 0-10) and manually annotated for each surgical action on every frame. The model was generated based on the training data. Validation of the model was performed on a set of unseen test data. Convolutional neural network (CNN)-based deep learning was also used.

RESULTS

The average surgical time was 175 min (± 43 min SD), with the individual surgical phases also showing high variations in the duration between cases. Each surgery started in the first phase (Phase 0) and ended in the last phase (Phase 10), and phase transitions occurred 14 (± 2 SD) times per procedure on an average. The accuracy of the automatic surgical phase recognition was 91.9% and those for the automatic surgical action recognition of extracorporeal action and irrigation were 89.4% and 82.5%, respectively. Moreover, this system could perform real-time automatic surgical phase recognition at 32 fps.

CONCLUSIONS

The CNN-based deep learning approach enabled the recognition of surgical phases and actions in 71 Lap-S cases based on manually annotated data. This system could perform automatic surgical phase recognition and automatic target surgical action recognition with high accuracy. Moreover, this study showed the feasibility of real-time automatic surgical phase recognition with high frame rate.

Using Surgeon Hand Motions to Predict Surgical Maneuvers

OBJECTIVE

This study explores how common machine learning techniques can predict surgical maneuvers from a continuous video record of surgical benchtop simulations.

BACKGROUND

Automatic computer vision recognition of surgical maneuvers (suturing, tying, and transition) could expedite video review and objective assessment of surgeries.

METHOD

We recorded hand movements of 37 clinicians performing simple and running subcuticular suturing benchtop simulations, and applied three machine learning techniques (decision trees, random forests, and hidden Markov models) to classify surgical maneuvers every 2 s (60 frames) of video.

RESULTS

Random forest predictions of surgical video correctly classified 74% of all video segments into suturing, tying, and transition states for a randomly selected test set. Hidden Markov model adjustments improved the random forest predictions to 79% for simple interrupted suturing on a subset of randomly selected participants.

CONCLUSION

Random forest predictions aided by hidden Markov modeling provided the best prediction of surgical maneuvers. Training of models across all users improved prediction accuracy by 10% compared with a random selection of participants.

APPLICATION

Marker-less video hand tracking can predict surgical maneuvers from a continuous video record with similar accuracy as robot-assisted surgical platforms, and may enable more efficient video review of surgical procedures for training and coaching.

Bioinspired Implementation and Assessment of a Remote-Controlled Robot

Daily activities are characterized by an increasing interaction with smart machines that present a certain level of autonomy. However, the intelligence of such electronic devices is not always transparent for the end user. This study is aimed at assessing the quality of the remote control of a mobile robot whether the artefact exhibits a human-like behavior or not. The bioinspired behavior implemented in the robot is the well-described two-thirds power law. The performance of participants who teleoperate the semiautonomous vehicle implementing the biological law is compared to a manual and nonbiological mode of control. The results show that the time required to complete the path and the number of collisions with obstacles are significantly lower in the biological condition than in the two other conditions. Also, the highest percentage of occurrences of curvilinear or smooth trajectories are obtained when the steering is assisted by an integration of the power law in the robot's way of working. This advanced analysis of the performance based on the naturalness of the movement kinematics provides a refined evaluation of the quality of the Human-Machine Interaction (HMI). This finding is consistent with the hypothesis of a relationship between the power law and jerk minimization. In addition, the outcome of this study supports the theory of a CNS origin of the power law. The discussion addresses the implications of the anthropocentric approach to enhance the HMI.

Automatic annotation of surgical activities using virtual reality environments

PURPOSE

Annotation of surgical activities becomes increasingly important for many recent applications such as surgical workflow analysis, surgical situation awareness, and the design of the operating room of the future, especially to train machine learning methods in order to develop intelligent assistance. Currently, annotation is mostly performed by observers with medical background and is incredibly costly and time-consuming, creating a major bottleneck for the above-mentioned technologies. In this paper, we propose a way to eliminate, or at least limit, the human intervention in the annotation process.

METHODS

Meaningful information about interaction between objects is inherently available in virtual reality environments. We propose a strategy to convert automatically this information into annotations in order to provide as output individual surgical process models.

VALIDATION

We implemented our approach through a peg-transfer task simulator and compared it to manual annotations. To assess the impact of our contribution, we studied both intra- and inter-observer variability.

RESULTS AND CONCLUSION

In average, manual annotations took more than 12 min for 1 min of video to achieve low-level physical activity annotation, whereas automatic annotation is achieved in less than a second for the same video period. We also demonstrated that manual annotation introduced mistakes as well as intra- and inter-observer variability that our method is able to suppress due to the high precision and reproducibility.

Surgical motion analysis using discriminative interpretable patterns

OBJECTIVE

The analysis of surgical motion has received a growing interest with the development of devices allowing their automatic capture. In this context, the use of advanced surgical training systems makes an automated assessment of surgical trainee possible. Automatic and quantitative evaluation of surgical skills is a very important step in improving surgical patient care.

MATERIAL AND METHOD

In this paper, we present an approach for the discovery and ranking of discriminative and interpretable patterns of surgical practice from recordings of surgical motions. A pattern is defined as a series of actions or events in the kinematic data that together are distinctive of a specific gesture or skill level. Our approach is based on the decomposition of continuous kinematic data into a set of overlapping gestures represented by strings (bag of words) for which we compute comparative numerical statistic (tf-idf) enabling the discriminative gesture discovery via its relative occurrence frequency.

RESULTS

We carried out experiments on three surgical motion datasets. The results show that the patterns identified by the proposed method can be used to accurately classify individual gestures, skill levels and surgical interfaces. We also present how the patterns provide a detailed feedback on the trainee skill assessment.

CONCLUSIONS

The proposed approach is an interesting addition to existing learning tools for surgery as it provides a way to obtain a feedback on which parts of an exercise have been used to classify the attempt as correct or incorrect.

Sequential surgical signatures in micro-suturing task

PURPOSE

Surgical processes are generally only studied by identifying differences in populations such as participants or level of expertise. But the similarity between this population is also important in understanding the process. We therefore proposed to study these two aspects.

METHODS

In this article, we show how similarities in process workflow within a population can be identified as sequential surgical signatures. To this purpose, we have proposed a pattern mining approach to identify these signatures.

VALIDATION

We validated our method with a data set composed of seventeen micro-surgical suturing tasks performed by four participants with two levels of expertise.

RESULTS

We identified sequential surgical signatures specific to each participant, shared between participants with and without the same level of expertise. These signatures are also able to perfectly define the level of expertise of the participant who performed a new micro-surgical suturing task. However, it is more complicated to determine who the participant is, and the method correctly determines this information in only 64% of cases.

CONCLUSION

We show for the first time the concept of sequential surgical signature. This new concept has the potential to further help to understand surgical procedures and provide useful knowledge to define future CAS systems.

Knowledge transfer for surgical activity prediction

PURPOSE

Lack of annotated training data hinders automatic recognition and prediction of surgical activities necessary for situation-aware operating rooms. We propose using knowledge transfer to compensate for data deficit and improve prediction.

METHODS

We used two approaches to extract and transfer surgical process knowledge. First, we encoded semantic information about surgical terms using word embedding. Secondly, we passed knowledge between different clinical datasets of neurosurgical procedures using transfer learning.

RESULTS

The combination of two methods provided 22% improvement of activity prediction. We also made several pertinent observations about surgical practices based on the results of the performed transfer.

CONCLUSION

Word embedding boosts learning process. Transfer learning was shown to be more effective than a simple combination of data, especially for less similar procedures.

Surgical skills: Can learning curves be computed from recordings of surgical activities?

PURPOSE

Surgery is one of the riskiest and most important medical acts that are performed today. The need to improve patient outcomes and surgeon training, and to reduce the costs of surgery, has motivated the equipment of operating rooms with sensors that record surgical interventions. The richness and complexity of the data that are collected call for new methods to support computer-assisted surgery. The aim of this paper is to support the monitoring of junior surgeons learning their surgical skill sets.

METHODS

Our method is fully automatic and takes as input a series of surgical interventions each represented by a low-level recording of all activities performed by the surgeon during the intervention (e.g., cut the skin with a scalpel). Our method produces a curve describing the process of standardization of the behavior of junior surgeons. Given the fact that junior surgeons receive constant feedback from senior surgeons during surgery, these curves can be directly interpreted as learning curves.

RESULTS

Our method is assessed using the behavior of a junior surgeon in anterior cervical discectomy and fusion surgery over his first three years after residency. They revealed the ability of the method to accurately represent the surgical skill evolution. We also showed that the learning curves can be computed by phases allowing a finer evaluation of the skill progression.

CONCLUSION

Preliminary results suggest that our approach constitutes a useful addition to surgical training monitoring.

Evaluation of robotic surgery skills using dynamic time warping

BACKGROUND AND OBJECTIVE

accompanied with the wide acceptance of robot assisted minimally invasive surgery (RMIS), the demand for efficient and objective surgical skills evaluation method is increased. Recently, with the development of medical engineering technology, several evaluation methods have been proposed. Among them, kinematic analysis, an unsupervised and data-based method, has been accepted by many researchers. However, this method is still limited by the number of metrics and unconvinced scoring system. This paper aims to propose a new evaluation method to assess surgical skills efficiently and objectively.

METHODS

this research proposed an efficient and effective surgical skills evaluation algorithm which used the trajectories of instrument tip and dynamic time warping (DTW) to provide trainees with real-time and summative feedback. The optimum trajectories based on 'Therbligs' theory was designed as a template. DTW algorithm was used to align actual trajectories to optimum trajectories with an evaluating indicator designed to emphasize the crucial motion features in surgical skills evaluation. The real-time feedback was obtained through a sliding time window to help trainees improve learning efficiency.

RESULTS

experts (n = 2) and novices (n = 8) were invited to complete the peg transfer tasks and 60 instrument tip trajectories were assessed by the proposed algorithm. Significant differences between different groups were observed (experts' right trajectories versus experts' left trajectories, p = 0.0002; experts' right trajectories versus novices' right trajectories, p = 0.0124). In addition, evaluation results of trajectories with operational mistakes were significantly different from those of others.

CONCLUSIONS

the proposed evaluation method showed its advantages in distinguishing and evaluating surgical performance. Given its ability to evaluate the performance based on kinematic information, the proposed evaluation method can be further developed in the future. Furthermore, because the proposed method can provide real-time feedback, it also has the potential to be a monitoring system in operation room.

Surgical data processing for smart intraoperative assistance systems

Different components of the newly defined field of surgical data science have been under research at our groups for more than a decade now. In this paper, we describe our sensor-driven approaches to workflow recognition without the need for explicit models, and our current aim is to apply this knowledge to enable context-aware surgical assistance systems, such as a unified surgical display and robotic assistance systems. The methods we evaluated over time include dynamic time warping, hidden Markov models, random forests, and recently deep neural networks, specifically convolutional neural networks.

Energy-Based Metrics for Arthroscopic Skills Assessment

Minimally invasive skills assessment methods are essential in developing efficient surgical simulators and implementing consistent skills evaluation. Although numerous methods have been investigated in the literature, there is still a need to further improve the accuracy of surgical skills assessment. Energy expenditure can be an indication of motor skills proficiency. The goals of this study are to develop objective metrics based on energy expenditure, normalize these metrics, and investigate classifying trainees using these metrics. To this end, different forms of energy consisting of mechanical energy and work were considered and their values were divided by the related value of an ideal performance to develop normalized metrics. These metrics were used as inputs for various machine learning algorithms including support vector machines (SVM) and neural networks (NNs) for classification. The accuracy of the combination of the normalized energy-based metrics with these classifiers was evaluated through a leave-one-subject-out cross-validation. The proposed method was validated using 26 subjects at two experience levels (novices and experts) in three arthroscopic tasks. The results showed that there are statistically significant differences between novices and experts for almost all of the normalized energy-based metrics. The accuracy of classification using SVM and NN methods was between 70% and 95% for the various tasks. The results show that the normalized energy-based metrics and their combination with SVM and NN classifiers are capable of providing accurate classification of trainees. The assessment method proposed in this study can enhance surgical training by providing appropriate feedback to trainees about their level of expertise and can be used in the evaluation of proficiency.

Temporal clustering of surgical activities in robot-assisted surgery

Purpose

Most evaluations of surgical workflow or surgeon skill use simple, descriptive statistics (e.g., time) across whole procedures, thereby deemphasizing critical steps and potentially obscuring critical inefficiencies or skill deficiencies. In this work, we examine off-line, temporal clustering methods that chunk training procedures into clinically relevant surgical tasks or steps during robot-assisted surgery.

Methods

We collected system kinematics and events data from nine surgeons performing five different surgical tasks on a porcine model using the da Vinci Si surgical system. The five tasks were treated as one ‘pseudo-procedure.’ We compared four different temporal clustering algorithms—hierarchical aligned cluster analysis (HACA), aligned cluster analysis (ACA), spectral clustering (SC), and Gaussian mixture model (GMM)—using multiple feature sets.

Results

HACA outperformed the other methods reaching an average segmentation accuracy of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$88.0\%$$\end{document}88.0% when using all system kinematics and events data as features. SC and ACA reached moderate performance with \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$84.1\%$$\end{document}84.1% and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$82.9\%$$\end{document}82.9% average segmentation accuracy, respectively. GMM consistently performed poorest across algorithms.

Conclusions

Unsupervised temporal segmentation of surgical procedures into clinically relevant steps achieves good accuracy using just system data. Such methods will enable surgeons to receive directed feedback on individual surgical tasks rather than whole procedures in order to improve workflow, assessment, and training.