Text segmentation is a challenging vision task with many downstream applications. Current text segmentation methods require pixel-level annotations, which are expensive in the cost of human labor and limited in application scenarios. In this paper, we take the first attempt to perform weakly-supervised text instance segmentation by bridging text recognition and text segmentation. The insight is that text recognition methods provide precise attention position of each text instance, and the attention location can feed to both a text adaptive refinement head (TAR) and a text segmentation head. Specifically, the proposed TAR generates pseudo labels by performing two-stage iterative refinement operations on the attention location to fit the accurate boundaries of the corresponding text instance. Meanwhile, the text segmentation head takes the rough attention location to predict segmentation masks which are supervised by the aforementioned pseudo labels. In addition, we design a mask-augmented contrastive learning by treating our segmentation result as an augmented version of the input text image, thus improving the visual representation and further enhancing the performance of both recognition and segmentation. The experimental results demonstrate that the proposed method significantly outperforms weakly-supervised instance segmentation methods on ICDAR13-FST (18.95$\%$ improvement) and TextSeg (17.80$\%$ improvement) benchmarks.
Heatmap-based methods play an important role in anatomical landmark detection. However, most current heatmap-based methods assume that the distributions of all landmarks are the same and the distribution of each landmark is isotropic, which may not be in line with reality. For example, the landmark on the jaw is more likely to be located along the edge and less likely to be located inside or outside the jaw. Manually annotating tends to follow similar rules, resulting in an anisotropic distribution for annotated landmarks, which represents the uncertainty in the annotation. To estimate the uncertainty, we propose a module named Pyramid Covariance Predictor to predict the covariance matrices of the target Gaussian distributions, which determine the distributions of landmarks and represent the uncertainty of landmark annotation. Specifically, the Pyramid Covariance Predictor utilizes the pyramid features extracted by the encoder of the backbone U-Net and predicts the Cholesky decomposition of the covariance matrix of the landmark location distribution. Experimental results show that the proposed Pyramid Covariance Predictor can accurately predict the distributions and improve the performance of anatomical landmark detection.
Subgraph-wise sampling -- a promising class of mini-batch training techniques for graph neural networks (GNNs -- is critical for real-world applications. During the message passing (MP) in GNNs, subgraph-wise sampling methods discard messages outside the mini-batches in backward passes to avoid the well-known neighbor explosion problem, i.e., the exponentially increasing dependencies of nodes with the number of MP iterations. However, discarding messages may sacrifice the gradient estimation accuracy, posing significant challenges to their convergence analysis and convergence speeds. To address this challenge, we propose a novel subgraph-wise sampling method with a convergence guarantee, namely Local Message Compensation (LMC). To the best of our knowledge, LMC is the first subgraph-wise sampling method with provable convergence. The key idea is to retrieve the discarded messages in backward passes based on a message passing formulation of backward passes. By efficient and effective compensations for the discarded messages in both forward and backward passes, LMC computes accurate mini-batch gradients and thus accelerates convergence. Moreover, LMC is applicable to various MP-based GNN architectures, including convolutional GNNs (finite message passing iterations with different layers) and recurrent GNNs (infinite message passing iterations with a shared layer). Experiments on large-scale benchmarks demonstrate that LMC is significantly faster than state-of-the-art subgraph-wise sampling methods.
For any video codecs, the coding efficiency highly relies on whether the current signal to be encoded can find the relevant contexts from the previous reconstructed signals. Traditional codec has verified more contexts bring substantial coding gain, but in a time-consuming manner. However, for the emerging neural video codec (NVC), its contexts are still limited, leading to low compression ratio. To boost NVC, this paper proposes increasing the context diversity in both temporal and spatial dimensions. First, we guide the model to learn hierarchical quality patterns across frames, which enriches long-term and yet high-quality temporal contexts. Furthermore, to tap the potential of optical flow-based coding framework, we introduce a group-based offset diversity where the cross-group interaction is proposed for better context mining. In addition, this paper also adopts a quadtree-based partition to increase spatial context diversity when encoding the latent representation in parallel. Experiments show that our codec obtains 23.5% bitrate saving over previous SOTA NVC. Better yet, our codec has surpassed the under-developing next generation traditional codec/ECM in both RGB and YUV420 colorspaces, in terms of PSNR. The codes are at https://github.com/microsoft/DCVC.
Needle picking is a challenging surgical task in robot-assisted surgery due to the characteristics of small slender shapes of needles, needles' variations in shapes and sizes, and demands for millimeter-level control. Prior works, heavily relying on the prior of needles (e.g., geometric models), are hard to scale to unseen needles' variations. In addition, visual tracking errors can not be minimized online using their approaches. In this paper, we propose an end-to-end deep visual learning framework for needle-picking tasks where both visual and control components can be learned jointly online. Our proposed framework integrates a state-of-the-art reinforcement learning framework, Dreamer, with behavior cloning (BC). Besides, two novel techniques, i.e., Virtual Clutch and Dynamic Spotlight Adaptation (DSA), are introduced to our end-to-end visual controller for needle-picking tasks. We conducted extensive experiments in simulation to evaluate the performance, robustness, variation adaptation, and effectiveness of individual components of our method. Our approach, trained by 8k demonstration timesteps and 140k online policy timesteps, can achieve a remarkable success rate of 80%, a new state-of-the-art with end-to-end vision-based surgical robot learning for delicate operations tasks. Furthermore, our method effectively demonstrated its superiority in generalization to unseen dynamic scenarios with needle variations and image disturbance, highlighting its robustness and versatility. Codes and videos are available at https://sites.google.com/view/dreamerbc.
Recent advancements toward perception and decision-making of flexible endoscopes have shown great potential in computer-aided surgical interventions. However, owing to modeling uncertainty and inter-patient anatomical variation in flexible endoscopy, the challenge remains for efficient and safe navigation in patient-specific scenarios. This paper presents a novel data-driven framework with self-contained visual-shape fusion for autonomous intelligent navigation of flexible endoscopes requiring no priori knowledge of system models and global environments. A learning-based adaptive visual servoing controller is proposed to online update the eye-in-hand vision-motor configuration and steer the endoscope, which is guided by monocular depth estimation via a vision transformer (ViT). To prevent unnecessary and excessive interactions with surrounding anatomy, an energy-motivated shape planning algorithm is introduced through entire endoscope 3-D proprioception from embedded fiber Bragg grating (FBG) sensors. Furthermore, a model predictive control (MPC) strategy is developed to minimize the elastic potential energy flow and simultaneously optimize the steering policy. Dedicated navigation experiments on a robotic-assisted flexible endoscope with an FBG fiber in several phantom environments demonstrate the effectiveness and adaptability of the proposed framework.
Task automation of surgical robot has the potentials to improve surgical efficiency. Recent reinforcement learning (RL) based approaches provide scalable solutions to surgical automation, but typically require extensive data collection to solve a task if no prior knowledge is given. This issue is known as the exploration challenge, which can be alleviated by providing expert demonstrations to an RL agent. Yet, how to make effective use of demonstration data to improve exploration efficiency still remains an open challenge. In this work, we introduce Demonstration-guided EXploration (DEX), an efficient reinforcement learning algorithm that aims to overcome the exploration problem with expert demonstrations for surgical automation. To effectively exploit demonstrations, our method estimates expert-like behaviors with higher values to facilitate productive interactions, and adopts non-parametric regression to enable such guidance at states unobserved in demonstration data. Extensive experiments on $10$ surgical manipulation tasks from SurRoL, a comprehensive surgical simulation platform, demonstrate significant improvements in the exploration efficiency and task success rates of our method. Moreover, we also deploy the learned policies to the da Vinci Research Kit (dVRK) platform to show the effectiveness on the real robot. Code is available at https://github.com/med-air/DEX.
Generalization in partially observed markov decision processes (POMDPs) is critical for successful applications of visual reinforcement learning (VRL) in real scenarios. A widely used idea is to learn task-relevant representations that encode task-relevant information of common features in POMDPs, i.e., rewards and transition dynamics. As transition dynamics in the latent state space -- which are task-relevant and invariant to visual distractions -- are unknown to the agents, existing methods alternatively use transition dynamics in the observation space to extract task-relevant information in transition dynamics. However, such transition dynamics in the observation space involve task-irrelevant visual distractions, degrading the generalization performance of VRL methods. To tackle this problem, we propose the reward sequence distribution conditioned on the starting observation and the predefined subsequent action sequence (RSD-OA). The appealing features of RSD-OA include that: (1) RSD-OA is invariant to visual distractions, as it is conditioned on the predefined subsequent action sequence without task-irrelevant information from transition dynamics, and (2) the reward sequence captures long-term task-relevant information in both rewards and transition dynamics. Experiments demonstrate that our representation learning approach based on RSD-OA significantly improves the generalization performance on unseen environments, outperforming several state-of-the-arts on DeepMind Control tasks with visual distractions.
With the high flexibility of supporting resource-intensive and time-sensitive applications, unmanned aerial vehicle (UAV)-assisted mobile edge computing (MEC) is proposed as an innovational paradigm to support the mobile users (MUs). As a promising technology, digital twin (DT) is capable of timely mapping the physical entities to virtual models, and reflecting the MEC network state in real-time. In this paper, we first propose an MEC network with multiple movable UAVs and one DT-empowered ground base station to enhance the MEC service for MUs. Considering the limited energy resource of both MUs and UAVs, we formulate an online problem of resource scheduling to minimize the weighted energy consumption of them. To tackle the difficulty of the combinational problem, we formulate it as a Markov decision process (MDP) with multiple types of agents. Since the proposed MDP has huge state space and action space, we propose a deep reinforcement learning approach based on multi-agent proximal policy optimization (MAPPO) with Beta distribution and attention mechanism to pursue the optimal computation offloading policy. Numerical results show that our proposed scheme is able to efficiently reduce the energy consumption and outperforms the benchmarks in performance, convergence speed and utilization of resources.