Surgical Gesture Research Articles

CLUSTERING SURGEMES USING PROTOTYPES FROM ROBOTIC KINEMATIC INFORMATION

Training a surgeon to be skilled and competent to perform a given surgical procedure is essential in providing a high quality of care and reducing the risk of complications. However, existing training techniques limit us from conducting in-depth analyses of surgical motions to evaluate these skills accurately. We develop a method to identify the gestures by applying unsupervised methods to cluster the surgical activities learned directly from raw kinematic data. We design an unsupervised method to determine the surgical motions in a Suturing procedure based on predefined surgical gestures. The first step is to find the prototypes by clustering the surgemes of the expert surgeon from all the same expert trials. Then, we map the other surgeons surgemes to the nearest representative of the prototypes and report the clustering accuracy by employing the rand index technique. We utilize four techniques in our proposed unsupervised approach for gesture clustering based on Hierarchical and FCM algorithms. In addition, we highlight the advantages of representing time series data before clustering in terms of computation time saving and system complexity reduction, respectively.

A Rare Case of a Left Ovarian Remnant Syndrome Associated with a Symptomatic Right Ovarian Cyst

Ovarian Remnant Syndrome (ORS) results from the presence of residual ovarian tissue after an ovariectomy. We report the case of a 39-year-old woman with a past history of total abdominal hysterectomy and left salpingooophorectomy who presented with pelvic pain and was diagnosed as having ORS. She was successfully operated by laparoscopy with the surgical gesture being the removal of the cyst and the surrounding ovarian tissue with the postoperative course being uneventful.

Gesture Recognition in Robotic Surgery With Multimodal Attention.

Automatically recognising surgical gestures from surgical data is an important building block of automated activity recognition and analytics, technical skill assessment, intra-operative assistance and eventually robotic automation. The complexity of articulated instrument trajectories and the inherent variability due to surgical style and patient anatomy make analysis and fine-grained segmentation of surgical motion patterns from robot kinematics alone very difficult. Surgical video provides crucial information from the surgical site with context for the kinematic data and the interaction between the instruments and tissue. Yet sensor fusion between the robot data and surgical video stream is non-trivial because the data have different frequency, dimensions and discriminative capability. In this paper, we integrate multimodal attention mechanisms in a two-stream temporal convolutional network to compute relevance scores and weight kinematic and visual feature representations dynamically in time, aiming to aid multimodal network training and achieve effective sensor fusion. We report the results of our system on the JIGSAWS benchmark dataset and on a new in vivo dataset of suturing segments from robotic prostatectomy procedures. Our results are promising and obtain multimodal prediction sequences with higher accuracy and better temporal structure than corresponding unimodal solutions. Visualization of attention scores also gives physically interpretable insights on network understanding of strengths and weaknesses of each sensor.

Surgesture: a novel instrument based on surgical actions for objective skill assessment

Due to varied surgical skills and the lack of an efficient rating system, we developed Surgesture based on elementary functional surgical gestures performed by surgeons, which could serve as objective metrics to evaluate surgical performance in laparoscopic cholecystectomy (LC). We defined 14 LC basic Surgestures. Four surgeons annotated Surgestures among LC videos performed by experts and novices. The counts, durations, average action time, and dissection/exposure ratio (D/E ratio) of LC Surgestures were compared. The phase of mobilizing hepatocystic triangle (MHT) was extracted for skill assessment by three professors using a modified Global Operative Assessment of Laparoscopic Skills (mGOALS). The novice operation time was significantly longer than the expert operation time (58.12 ± 19.23min vs. 26.66 ± 8.00min, P < 0.001), particularly during MHT phase. Novices had significantly more Surgestures than experts in both hands (P < 0.05). The left hand and inefficient Surgesture of novices were dramatically more than those of experts (P < 0.05). The experts demonstrated a significantly higher D/E ratio of duration than novices (0.79 ± 0.37 vs. 2.84 ± 1.98, P < 0.001). The counts and time pattern map of LC Surgestures during MHT demonstrated that novices tended to complete LC with more types of Surgestures and spent more time exposing the surgical scene. The performance metrics of LC Surgesture had significant but weak associations with each aspect of mGOALS. The newly constructed Surgestures could serve as accessible and quantifiable metrics for demonstrating the operative pattern and distinguishing surgeons with various skills. The association between Surgestures and Global Rating Scale laid the foundation for establishing a bridge to automated objective surgical skill evaluation.

Using open surgery simulation kinematic data for tool and gesture recognition.

The use of motion sensors is emerging as a means for measuring surgical performance. Motion sensors are typically used for calculating performance metrics and assessing skill. The aim of this study was to identify surgical gestures and tools used during an open surgery suturing simulation based on motion sensor data. Twenty-five participants performed a suturing task on a variable tissue simulator. Electromagnetic motion sensors were used to measure their performance. The current study compares GRU and LSTM networks, which are known to perform well on other kinematic datasets, as well as MS-TCN++, which was developed for video data and was adapted in this work for motion sensors data. Finally, we extended all architectures for multi-tasking. In the gesture recognition task the MS-TCN++ has the highest performance with accuracy of [Formula: see text] and F1-Macro of [Formula: see text], edit distance of [Formula: see text] and F1@10 of [Formula: see text] In the tool usage recognition task for the right hand, MS-TCN++ performs the best in most metrics with an accuracy score of [Formula: see text], F1-Macro of [Formula: see text], F1@10 of [Formula: see text], and F1@25 of [Formula: see text]. The multi-task GRU performs best in all metrics in the left-hand case, with an accuracy of [Formula: see text], edit distance of [Formula: see text], F1-Macro of [Formula: see text], F1@10 of [Formula: see text], and F1@25 of [Formula: see text]. In this study, using motion sensor data, we automatically identified the surgical gestures and the tools used during an open surgery suturing simulation. Our methods may be used for computing more detailed performance metrics and assisting in automatic workflow analysis. MS-TCN++ performed better in gesture recognition as well as right-hand tool recognition, while the multi-task GRU provided better results in the left-hand case. It should be noted that our multi-task GRU network is significantly smaller and has achieved competitive results in the rest of the tasks as well.

Telestration in the Teaching of Basic Surgical Skills: A Randomized Trial

To evaluate how an affordable course using telestration with augmented reality can be compared to the traditional teaching of basic surgical skills. Prospective, randomized and blinded study. Faculty of Medicine of PortoUniversity. Twenty medical students without any experience in basic surgical skills were randomized into two different learning groups: telestration and traditional teaching (on-site mentoring) groups. Five different types of sutures were taught: the single interrupted, the cruciate mattress, the horizontal mattress, the vertical mattress and the simple continuous sutures. Data was obtained on the time taken to learn each of the techniques and to perform each exercise without any support from the faculty, tension of the suture, quality of the procedure using a modified Objective Structured Assessment of Technical Skills and participants' answers to a Likert questionnaire in terms of their learning experience, confidence, and self-evaluation. Trainees in the telestration group were globally faster when performing independently (1393.40 [SD 288.89] vs 1679.00 [SD 328.22] seconds, p = 0.04) particularly during the cruciate mattress suture (235.50 [SD 61.81] vs 290.00 [SD 68.77] seconds, p = 0.05) and the simple continuous suture (492.40 [SD 87.49] vs 630.30 [SD 132.34] seconds, p = 0.01).Time needed for students to learn the procedures was similar between the groups. There were also no statistically significant differences in terms of the quality of the surgical gesture, tension of the suture, self-evaluation or confidence. A basic surgical skills course using telestration through a head-mounted device with augmented reality capabilities can be a viable alternative to traditional teaching, considering time and quality of the gesture. Though costs can discourage from using this technology in basic procedures, the use of free software may turn it into an affordable option in the context of distant learning.

Generalization of Deep Learning Gesture Classification in Robotic-Assisted Surgical Data: From Dry Lab to Clinical-Like Data.

Robotic-assisted minimally invasive surgery (RAMIS) became a common practice in modern medicine and is widely studied. Surgical procedures require prolonged and complex movements; therefore, classifying surgical gestures could be helpful to characterize surgeon performance. The public release of the JIGSAWS dataset facilitates the development of classification algorithms; however, it is not known how algorithms trained on dry-lab data generalize to real surgical situations. We trained a Long Short-Term Memory (LSTM) network for the classification of dry lab and clinical-like data into gestures. We show that a network that was trained on the JIGSAWS data does not generalize well to other dry-lab data and to clinical-like data. Using rotation augmentation improves performance on dry-lab tasks, but fails to improve the performance on clinical-like data. However, using the same network architecture, adding the six joint angles of the patient-side manipulators (PSMs) features, and training the network on the clinical-like data together lead to notable improvement in the classification of the clinical-like data. Using the JIGSAWS dataset alone is insufficient for training a gesture classification network for clinical data. However, it can be very informative for determining the architecture of the network, and with training on a small sample of clinical data, can lead to acceptable classification performance. Developing efficient algorithms for gesture classification in clinical surgical data is expected to advance understanding of surgeon sensorimotor control in RAMIS, the automation of surgical skill evaluation, and the automation of surgery.

Recurrent Prosthetic Knee Infection Associated with a Hidden Nidus in the Quadriceps Tendon

We present a case of a 60-year-old male with poorly controlled type 2 Diabetes that had recurrent prosthetic knee infection that besides antibiotic or surgical treatment was difficult to treat. On the latest revision, he presented with an infection of the knee spacer with severe stiffness impairing knee flexion. In the debridement phase, we noticed a nidus of purulent tissue on the quadriceps tendon hidden inside an apparently normal peritendon. This case highlights several management details to avoid recurrent catastrophic prosthetic knee infection recurrence such as: the importance of perioperative management of comorbid diseases such as hyperglycaemia and poorly controlled diabetes that are a well-defined evidence-based risk factor for prosthetic joint infection and also the importance of adequate debridement as a balanced surgical gesture between excising too little or too much not to damage the soft tissue envelope.

CHIVA cure and sclerotherapy

Sclerotherapy is, by definition, a destructive treatment and, as such, it is conceptually long far away from the basic principles of the conservative hemodynamic cure for venous insufficiency (CHIVA cure). Accordingly, Phlebologists and Vascular Surgeons who practice the CHIVA cure at most acknowledge the use of sclerotherapy after the CHIVA cure for the aesthetical refinement or for the control over time of small calibre recurrences. However, intraoperative sclerotherapy can be used in association with the CHIVA cure in those cases in which the surgical gesture foreseen by the CHIVA cure would not be able to flush disconnect escape points at the level of the deep venous system and, as a consequence, it would leave non-draining stumps, which are a source of recurrences. Further, under some particular circumstances, sclerotherapy can be used instead of the CHIVA cure for the treatment of type II shunts or as the 1st step of the CHIVA 2 strategy for the treatment of type III shunts. This paper describes the use of intraoperative sclerotherapy for the treatment of escape points from the deep venous system difficult to treat by surgery, such as perforator veins, some anatomic presentations of the saphenous-popliteal junction and some pelvic shunts. This paper also describes the use of sclerotherapy alone for the treatment of type II shunts or for modelling saphenous hemodynamics in type III shunts before completing the treatment by surgery.

Robot-Assisted Middle Ear Endoscopic Surgery: Preliminary Results on 37 Patients.

Background: Endoscopy during middle ear surgery is advantageous for better exploration of middle ear structures. However, using an endoscope has some weaknesses as surgical gestures are performed with one hand. This may trouble surgeons accustomed to using two-handed surgery, and may affect accuracy. A robot-based holder may combine the benefits from endoscopic exposure with a two-handed technique. The purpose of this study was to assess the safety and value of an endoscope held by a teleoperated system.Patients and Methods: A case series of 37 consecutive patients operated using endoscopic exposure with robot-based assistance was analyzed retrospectively. The RobOtol® system (Collin, France) was teleoperated as an endoscope holder in combination with a microscope. The following data were collected: patient characteristics, etiology, procedure type, complications, mean air and bone conduction thresholds, and speech performance at 3 months postoperatively. Patients had type I (myringoplasty), II (partial ossiculoplasty), and III (total ossiculoplasty) tympanoplasties in 15, 14, and 4 cases, respectively. Three patients had partial petrosectomies for cholesteatomas extending to the petrous apex. Finally, one case underwent resection of a tympanic paraganglioma. Ambulatory procedures were performed in 25 of the 37 patients (68%).Results: Complete healing with no perforation of the tympanic membrane was noted postoperatively in all patients. No complications relating to robotic manipulation occurred during surgery or postoperatively. The mean air conduction gain was 3.8 ± 12.6 dB for type I (n = 15), 7.9 ± 11.4 dB for type II (n = 14), and −0.9 ± 10.8 for type III tympanoplasties (n = 4), and the postoperative air-bone conduction gap was 13.8 ± 13.3 dB for type I, 19.7 ± 11.7 dB for type II and 31.6 ± 13.0 dB for type III tympanoplasty. They was no relapse of cholesteatoma or paraganglioma during the short follow-up period (<1 year).Conclusion: This study indicates that robot-assisted endoscopy is a safe and trustworthy tool for several categories of middle ear procedures. It combines the benefits of endoscopic exposure with a two-handed technique in middle ear surgery. It can be used as a standalone tool for pathology limited to the middle ear cleft or in combination with a microscope in lesions extending to the mastoid or petrous apex.

Minimally Invasive Neck Surgery: An Animal Model Study.

Background: Minimally invasive surgery (MIS) is replacing conventional surgery as the "gold standard" in different surgical areas. Although cervical MIS is already accepted in the adult population, its use in children is still new and controversial. The natural obstacles to this approach are the absence of a natural cavity, with the inherent complications of creating one artificially, and the limited existing workspace especially in pediatric patients. All endoscopic techniques in the field of neck surgery try to live up to the high cosmetic expectations and the transoral cervical approach as a natural orifice surgery technique excels at it. Aim: Besides the goal of feasibility, we aim to report on the pitfalls of this approach, by using an experimental rabbit model for minimally invasive thyroidectomy. Materials and Methods: Transoral endoscopic thyroidectomies using a vestibular approach were performed in 10 anesthetized rabbits. All surgeries were video recorded. The surgical time, anatomy identified, difficulties, and intraoperative complications were documented. Results: Through one trocar in the vestibular area and two lateral stab incisions, it was possible to create a working space and to reach the peritracheal area. Total thyroidectomies were completed in the 10 animals with a mean operative time of 51 minutes. In all of them we were able to identify the fascial spaces of the neck and the major vessels. During surgery, the lack of space required surgical gestures to be very precise, soft, and gentle. There were 2 cases with a small amount of bleeding and one mild trachea laceration during the procedure, but none of them required suspension or conversion to an open procedure. Animals were euthanized immediately after the surgical procedure. Conclusions: The vestibular approach seems to be a feasible technique to access pediatric neck pathology. Despite the differences in the cervical anatomy, the limited workspace of the rabbit model perfectly matches the requirements of a pediatric training model.

SD-Net: joint surgical gesture recognition and skill assessment

PurposeSurgical gesture recognition has been an essential task for providing intraoperative context-aware assistance and scheduling clinical resources. However, previous methods present limitations in catching long-range temporal information, and many of them require additional sensors. To address these challenges, we propose a symmetric dilated network, namely SD-Net, to jointly recognize surgical gestures and assess surgical skill levels only using RGB surgical video sequences.MethodsWe utilize symmetric 1D temporal dilated convolution layers to hierarchically capture gesture clues under different receptive fields such that features in different time span can be aggregated. In addition, a self-attention network is bridged in the middle to calculate the global frame-to-frame relativity.ResultsWe evaluate our method on a robotic suturing task from the JIGSAWS dataset. The gesture recognition task largely outperforms the state of the arts on the frame-wise accuracy up to sim 6 points and the F1@50 score sim 8 points. We also keep the 100% predicted accuracy for the skill assessment task using LOSO validation scheme.ConclusionThe results indicate that our architecture is able to obtain representative surgical video features by extensively considering the spatial, temporal and relational context from raw video input. Furthermore, the better performance in multi-task learning implies that surgical skill assessment has a complementary effects to gesture recognition task.

A Technique for a More Accurate Acromioclavicular Joint Reduction in Arthroscopic Coracoclavicular Stabilization of Acromioclavicular Joint Dislocation

AbstractLoss of reduction is the most common complication following acromioclavicular dislocations treatment, with literature showing greater postoperative coracoclavicular distances associated with worse clinical results. We present a surgical gesture that aims to help surgeons achieve and secure an anatomic acromioclavicular reduction during coracoclavicular fixation. This technique has the possibility to improve radiological and functional results of acromioclavicular dislocation treatment.

Cross-modal self-supervised representation learning for gesture and skill recognition in robotic surgery.

Multi- and cross-modal learning consolidates information from multiple data sources which may offer a holistic representation of complex scenarios. Cross-modal learning is particularly interesting, because synchronized data streams are immediately useful as self-supervisory signals. The prospect of achieving self-supervised continual learning in surgical robotics is exciting as it may enable lifelong learning that adapts to different surgeons and cases, ultimately leading to a more general machine understanding of surgical processes. We present a learning paradigm using synchronous video and kinematics from robot-mediated surgery. Our approach relies on an encoder-decoder network that maps optical flow to the corresponding kinematics sequence. Clustering on the latent representations reveals meaningful groupings for surgeon gesture and skill level. We demonstrate the generalizability of the representations on the JIGSAWS dataset by classifying skill and gestures on tasks not used for training. For tasks seen in training, we report a 59 to 70% accuracy in surgical gestures classification. On tasks beyond the training setup, we note a 45 to 65% accuracy. Qualitatively, we find that unseen gestures form clusters in the latent space of novice actions, which may enable the automatic identification of novel interactions in a lifelong learning scenario. From predicting the synchronous kinematics sequence, optical flow representations of surgical scenes emerge that separate well even for new tasks that the model had not seen before. While the representations are useful immediately for a variety of tasks, the self-supervised learning paradigm may enable research in lifelong and user-specific learning.

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%.

Gesture Recognition in Robotic Surgery: A Review.

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

From the Dexterous Surgical Skill to the Battlefield—A Robotics Exploratory Study

Short response time is critical for future military medical operations in austere settings or remote areas. Such effective patient care at the point of injury can greatly benefit from the integration of semi-autonomous robotic systems. To achieve autonomy, robots would require massive libraries of maneuvers collected with the goal of training machine learning algorithms. Although this is attainable in controlled settings, obtaining surgical data in austere settings can be difficult. Hence, in this article, we present the Dexterous Surgical Skill (DESK) database for knowledge transfer between robots. The peg transfer task was selected as it is one of the six main tasks of laparoscopic training. In addition, we provide a machine learning framework to evaluate novel transfer learning methodologies on this database. A set of surgical gestures was collected for a peg transfer task, composed of seven atomic maneuvers referred to as surgemes. The collected Dexterous Surgical Skill dataset comprises a set of surgical robotic skills using the four robotic platforms: Taurus II, simulated Taurus II, YuMi, and the da Vinci Research Kit. Then, we explored two different learning scenarios: no-transfer and domain-transfer. In the no-transfer scenario, the training and testing data were obtained from the same domain; whereas in the domain-transfer scenario, the training data are a blend of simulated and real robot data, which are tested on a real robot. Using simulation data to train the learning algorithms enhances the performance on the real robot where limited or no real data are available. The transfer model showed an accuracy of 81% for the YuMi robot when the ratio of real-tosimulated data were 22% to 78%. For the Taurus II and the da Vinci, the model showed an accuracy of 97.5% and 93%, respectively, training only with simulation data. The results indicate that simulation can be used to augment training data to enhance the performance of learned models in real scenarios. This shows potential for the future use of surgical data from the operating room in deployable surgical robots in remote areas.

Rare localization of synovial cyst of the retroachilles bursa

Synovial cyst of the retroachilles bursa is a cystic formation developing from the retroachilles bursa. It is a rare cause of talalgia which has been very little reported in the literature. We report the case of a 22-year-old woman suffering from rebel left talalgia resistant to medical treatment. Left foot X-ray showed sharp and protruding bony outgrowth above and behind the great calcanean tuberosity. Chauveaux-Liet angle was 16 degrees and the Fowler-Philip angle was 82 degrees. The diagnosis of Haglund's deformity was retained. Dorsal surgery was performed. The cyst of the retroachilles bursa was removed and sent to the laboratory for anatomopathological examination. This confirmed the diagnosis of synovial cyst. The surgical gesture was complemented by removal of bony outgrowth. Patient's clinical outcome was favorable. This study reports a rare and unusual case of synovial cyst in an extra-articular location, complicating Haglund's deformity.

Role of uterine manipulator during laparoscopic endometrial cancer treatment.

Endometrial cancer (EC) is the most frequent gynecologic malignancy with 61,000 new cases diagnosed per year in the USA. Gynecologic Oncology Group (GOG) LAP2 randomized study demonstrated a superimposable 5-year overall survival between laparotomic and laparoscopic approaches in EC treatment. In this context the uterine manipulator (UM) represents a valuable tool in order to increase the ergonomics of surgical gesture during a laparoscopic total hysterectomy. Despite the proven safety of the minimally invasive approach in EC treatment, neither the indication whether to use or not, nor the recommendation concerning a specific type of UM are provided by international guidelines. This narrative review aims to collect all the main findings in the literature about UM use to investigate its safety in EC patients. Based on the analysis of the literature research, the main results were categorized into two manipulator-related problems: the iatrogenic LVSI, and the retrograde tumor spillage. LVSI is defined as the presence of tumor cells within an endothelium-lined space. An electronic search was performed using the following keywords: ‘uterine manipulator’, and ‘endometrial cancer’. The electronic database search provided a total of 93 studies. Of whom, 12 case reports, 5 studies not in English language, and 65 works not fitting the review scope were excluded from the analysis. Eleven studies were considered eligible for the purpose of the study. The most recent studies have highlighted the safety of the uterine manipulator in the early-stage EC laparoscopic treatment. All types of manipulators are considered to be fairly safe but its application should be tailored according to tumor dimension and grade of myometrial infiltration.

Surgical Hand Gesture Prediction for the Operating Room.

Technological advancements in smart assistive technology enable navigating and manipulating various types of computer-aided devices in the operating room through a contactless gesture interface. Understanding surgeon actions is crucial to natural human-robot interaction in operating room since it means a sort of prediction a human behavior so that the robot can foresee the surgeon's intention, early choose appropriate action and reduce waiting time. In this paper, we present a new deep network based on Convolution Long Short-Term Memory (ConvLSTM) for gesture prediction configured to provide natural interaction between the surgeon and assistive robot and improve operating-room efficiency. The experimental results prove the capability of reliably recognizing unfinished gestures on videos. We quantitatively demonstrate the latter ability and the fact that GestureConvLSTM improves the baseline system performance on LSA64 dataset.