Surgical context inference has recently garnered significant attention in robot-assisted surgery as it can facilitate workflow analysis, skill assessment, and error detection. However, runtime context inference is challenging since it requires timely and accurate detection of the interactions among the tools and objects in the surgical scene based on the segmentation of video data. On the other hand, existing state-of-the-art video segmentation methods are often biased against infrequent classes and fail to provide temporal consistency for segmented masks. This can negatively impact the context inference and accurate detection of critical states. In this study, we propose a solution to these challenges using a Space Time Correspondence Network (STCN). STCN is a memory network that performs binary segmentation and minimizes the effects of class imbalance. The use of a memory bank in STCN allows for the utilization of past image and segmentation information, thereby ensuring consistency of the masks. Our experiments using the publicly available JIGSAWS dataset demonstrate that STCN achieves superior segmentation performance for objects that are difficult to segment, such as needle and thread, and improves context inference compared to the state-of-the-art. We also demonstrate that segmentation and context inference can be performed at runtime without compromising performance.
Fine-grained activity recognition enables explainable analysis of procedures for skill assessment, autonomy, and error detection in robot-assisted surgery. However, existing recognition models suffer from the limited availability of annotated datasets with both kinematic and video data and an inability to generalize to unseen subjects and tasks. Kinematic data from the surgical robot is particularly critical for safety monitoring and autonomy, as it is unaffected by common camera issues such as occlusions and lens contamination. We leverage an aggregated dataset of six dry-lab surgical tasks from a total of 28 subjects to train activity recognition models at the gesture and motion primitive (MP) levels and for separate robotic arms using only kinematic data. The models are evaluated using the LOUO (Leave-One-User-Out) and our proposed LOTO (Leave-One-Task-Out) cross validation methods to assess their ability to generalize to unseen users and tasks respectively. Gesture recognition models achieve higher accuracies and edit scores than MP recognition models. But, using MPs enables the training of models that can generalize better to unseen tasks. Also, higher MP recognition accuracy can be achieved by training separate models for the left and right robot arms. For task-generalization, MP recognition models perform best if trained on similar tasks and/or tasks from the same dataset.
Phase retrieval (PR) is an essential problem in a number of coherent imaging systems. This work aims at resolving the holographic phase retrieval problem in real world scenarios where the measurements are corrupted by a mixture of Poisson and Gaussian (PG) noise that stems from optical imaging systems. To solve this problem, we develop a novel algorithm based on Accelerated Wirtinger Flow that uses Score-based Diffusion models as the generative prior (AWFSD). In particular, we frame the PR problem as an optimization task that involves both a data fidelity term and a regularization term. We derive the gradient of the PG log-likelihood function along with its corresponding Lipschitz constant, ensuring a more accurate data consistency term for practical measurements. We introduce a generative prior as part of our regularization approach by using a score-based diffusion model to capture (the gradient of) the image prior distribution. We provide theoretical analysis that establishes a critical-point convergence guarantee for the proposed AWFSD algorithm. Our simulation experiments demonstrate that: 1) The proposed algorithm based on the PG likelihood model enhances reconstruction compared to that solely based on either Gaussian or Poisson likelihood. 2) The proposed AWFSD algorithm produces reconstructions with higher image quality both qualitatively and quantitatively, and is more robust to variations in noise levels when compared with state-of-the-art methods for phase retrieval.
Manual labeling of gestures in robot-assisted surgery is labor intensive, prone to errors, and requires expertise or training. We propose a method for automated and explainable generation of gesture transcripts that leverages the abundance of data for image segmentation. Surgical context is detected using segmentation masks by examining the distances and intersections between the tools and objects. Next, context labels are translated into gesture transcripts using knowledge-based Finite State Machine (FSM) and data-driven Long Short Term Memory (LSTM) models. We evaluate the performance of each stage of our method by comparing the results with the ground truth segmentation masks, the consensus context labels, and the gesture labels in the JIGSAWS dataset. Our results show that our segmentation models achieve state-of-the-art performance in recognizing needle and thread in Suturing and we can automatically detect important surgical states with high agreement with crowd-sourced labels (e.g., contact between graspers and objects in Suturing). We also find that the FSM models are more robust to poor segmentation and labeling performance than LSTMs. Our proposed method can significantly shorten the gesture labeling process (~2.8 times).
Training end-to-end unrolled iterative neural networks for SPECT image reconstruction requires a memory-efficient forward-backward projector for efficient backpropagation. This paper describes an open-source, high performance Julia implementation of a SPECT forward-backward projector that supports memory-efficient backpropagation with an exact adjoint. Our Julia projector uses only ~5% of the memory of an existing Matlab-based projector. We compare unrolling a CNN-regularized expectation-maximization (EM) algorithm with end-to-end training using our Julia projector with other training methods such as gradient truncation (ignoring gradients involving the projector) and sequential training, using XCAT phantoms and virtual patient (VP) phantoms generated from SIMIND Monte Carlo (MC) simulations. Simulation results with two different radionuclides (90Y and 177Lu) show that: 1) For 177Lu XCAT phantoms and 90Y VP phantoms, training unrolled EM algorithm in end-to-end fashion with our Julia projector yields the best reconstruction quality compared to other training methods and OSEM, both qualitatively and quantitatively. For VP phantoms with 177Lu radionuclide, the reconstructed images using end-to-end training are in higher quality than using sequential training and OSEM, but are comparable with using gradient truncation. We also find there exists a trade-off between computational cost and reconstruction accuracy for different training methods. End-to-end training has the highest accuracy because the correct gradient is used in backpropagation; sequential training yields worse reconstruction accuracy, but is significantly faster and uses much less memory.
The concept of causality plays an important role in human cognition . In the past few decades, causal inference has been well developed in many fields, such as computer science, medicine, economics, and education. With the advancement of deep learning techniques, it has been increasingly used in causal inference against counterfactual data. Typically, deep causal models map the characteristics of covariates to a representation space and then design various objective optimization functions to estimate counterfactual data unbiasedly based on the different optimization methods. This paper focuses on the survey of the deep causal models, and its core contributions are as follows: 1) we provide relevant metrics under multiple treatments and continuous-dose treatment; 2) we incorporate a comprehensive overview of deep causal models from both temporal development and method classification perspectives; 3) we assist a detailed and comprehensive classification and analysis of relevant datasets and source code.
Objective: We propose a formal framework for modeling surgical tasks using a unified set of motion primitives (MPs) as the basic surgical actions to enable more objective labeling and aggregation of different datasets and training generalized models for surgical action recognition. Methods: We use our framework to create the COntext and Motion Primitive Aggregate Surgical Set (COMPASS), including six dry-lab surgical tasks from three publicly-available datasets (JIGSAWS, DESK, and ROSMA) with kinematic and video data and context and MP labels. Methods for labeling surgical context and automatic translation to MPs are presented. We propose the Leave-One-Task-Out (LOTO) cross validation method to evaluate a model's ability to generalize to an unseen task. Results: Our context labeling method achieves near-perfect agreement between consensus labels from crowd-sourcing and expert surgeons. Segmentation of tasks to MPs enables the generation of separate left and right transcripts and significantly improves LOTO performance. We find that MP segmentation models perform best if trained on tasks with the same context and/or tasks from the same dataset. Conclusion: The proposed framework enables high-quality labeling of surgical data based on context and fine-grained MPs. Modeling surgical tasks with MPs enables the aggregation of different datasets for training action recognition models that can generalize better to unseen tasks than models trained at the gesture level. Significance: Our formal framework and aggregate dataset can support the development of models and algorithms for surgical process analysis, skill assessment, error detection, and autonomy.
Despite significant developments in the design of surgical robots and automated techniques for objective evaluation of surgical skills, there are still challenges in ensuring safety in robot-assisted minimally-invasive surgery (RMIS). This paper presents a runtime monitoring system for the detection of executional errors during surgical tasks through the analysis of kinematic data. The proposed system incorporates dual Siamese neural networks and knowledge of surgical context, including surgical tasks and gestures, their distributional similarities, and common error modes, to learn the differences between normal and erroneous surgical trajectories from small training datasets. We evaluate the performance of the error detection using Siamese networks compared to single CNN and LSTM networks trained with different levels of contextual knowledge and training data, using the dry-lab demonstrations of the Suturing and Needle Passing tasks from the JIGSAWS dataset. Our results show that gesture specific task nonspecific Siamese networks obtain micro F1 scores of 0.94 (Siamese-CNN) and 0.95 (Siamese-LSTM), and perform better than single CNN (0.86) and LSTM (0.87) networks. These Siamese networks also outperform gesture nonspecific task specific Siamese-CNN and Siamese-LSTM models for Suturing and Needle Passing.
In this paper, we introduce a novel iterative algorithm for the problem of phase-retrieval where the measurements consist of only the magnitude of linear function of the unknown signal, and the noise in the measurements follow Poisson distribution. The proposed algorithm is based on the principle of majorization-minimization (MM); however, the application of MM here is very novel and distinct from the way MM has been usually used to solve optimization problems in the literature. More precisely, we reformulate the original minimization problem into a saddle point problem by invoking Fenchel dual representation of the log (.) term in the Poisson likelihood function. We then propose tighter surrogate functions over both primal and dual variables resulting in a double-loop MM algorithm, which we have named as Primal-Dual Majorization-Minimization (PDMM) algorithm. The iterative steps of the resulting algorithm are simple to implement and involve only computing matrix vector products. We also extend our algorithm to handle various L1 regularized Poisson phase-retrieval problems (which exploit sparsity). The proposed algorithm is compared with previously proposed algorithms such as wirtinger flow (WF), MM (conventional), and alternating direction methods of multipliers (ADMM) for the Poisson data model. The simulation results under different experimental settings show that PDMM is faster than the competing methods, and its performance in recovering the original signal is at par with the state-of-the-art algorithms.