Imitation learning (IL) has shown immense promise in enabling autonomous dexterous manipulations, including in learning surgical tasks. To fully unlock the potential of IL for surgery, access to clinical datasets is needed, which unfortunately lack the kinematic data required for current IL approaches. A promising source of large-scale surgical demonstrations is monocular surgical videos available online, making monocular pose estimation a crucial step toward enabling large-scale robot learning. Towards this end, we propose SurgiPose, a differentiable rendering-based approach to estimate kinematic information from monocular surgical videos, eliminating the need for direct access to ground-truth kinematics. Our method infers tool trajectories and joint angles by optimizing tool pose parameters to minimize the discrepancy between rendered and real images. To evaluate the effectiveness of our approach, we conduct experiments on two robotic surgical tasks—tissue lifting and needle pickup—using the da Vinci Research Kit Si (dVRK Si). We train imitation learning policies with both ground-truth measured kinematics and with estimated kinematics from video and compare their performance. Our results show that policies trained on estimated kinematics achieve comparable success rates to those trained on ground-truth data, demonstrating the feasibility of using monocular video-based kinematic estimation for surgical robot learning. By enabling kinematic estimation from monocular surgical videos, our work lays the foundation for large-scale learning of autonomous surgical policies from online surgical data.

We study the problem of extracting emotions and the causes behind these emotions in conversations. Existing methods either tackle them separately or jointly model them at the coarse-grained level of emotions (fewer emotion categories) and causes (utterance-level causes). In this work, we aim to jointly extract more fine-grained emotions and causes. We construct a fine-grained dataset FG-RECCON, includes 16 fine-grained emotion categories and span-level causes. To further improve the fine-grained extraction performance, we propose to utilize the casual discourse knowledge in a knowledge distillation way. Specifically, the teacher model learns to predict causal connective words between utterances, and then guides the student model in identifying both the fine-grained emotion labels and causal spans. Experimental results demonstrate that our distillation method achieves state-of-the-art performance on both RECCON and FG-RECCON dataset.

The multigrid (MG) method has been widely used for large-scale electromagnetic (EM) forward modeling, primarily due to its linear dependence between computational time and grid size. However, in natural source EM forward modeling, the inclusion of air layers and the use of low frequencies can result in slow convergence or even divergence for MG solvers, due to the sharp change of conductivity and violation of the divergence-free condition for current. One typical solution is to develop an efficient and specialized smoother, for instance, with one property of satisfying the divergence-free condition in certain local areas. In this article, we modify the curl-curl differential equation by explicitly including the gradient of the divergence of current (CCGD). Hence, this ensures approximately the divergence-free condition at every iteration while solving EM fields. Based on the modified equation, an MG solver is then developed with traditional Krylov subspace solvers as the smoother, and no specialized smoothers are required. One benchmark model, one reservoir model with topography, and one realistic model from inversion are used to examine its numerical performance, in terms of accuracy, efficiency, and ability to handle grid stretching and size increase.

In this paper, an efficient moving object segmentation algorithm in the wavelet domain is proposed using three successive frames. The change detection method, which employs fuzzy C-means clustering technique to classify motion features of four wavelet sub-bands, is used twice to separate significant change pixels in the wavelet domain from the background. After applying the intersect operation, the change detection masks are obtained in wavelet domain. Finally, further object shape information and accurate extraction of the moving object is obtained in original resolution according to current object edge map. The experimental results demonstrate the algorithm effective.