Search

Underwater images often suffer from various degradations, such as color distortion, reduced visibility, and uneven illumination, caused by light absorption, scattering, and artificial lighting. However, most existing methods have focused on addressing singular or dual degradation aspects, lacking a comprehensive solution to underwater image degradation. This limitation hinders the application of vision technology in underwater scenarios. In this paper, we propose a framework for enhancing the quality of multi-degraded underwater images. This framework is distinctive in its ability to concurrently address color degradation, hazy blur, and non-uniform illumination by fusing RGB and Event signals. Specifically, an adaptive underwater color compensation algorithm is first proposed, informed by an analysis of the color degradation characteristics prevalent in underwater images. This compensation algorithm is subsequently integrated with a white balance algorithm to achieve color correction. Then, a dehazing method is developed, leveraging the fusion of sharpened images and gamma-corrected images to restore blurry details in RGB images and event reconstruction images. Finally, an illumination map is extracted from the RGB image, and a multi-scale fusion strategy is employed to merge the illumination map with the event reconstruction image, effectively enhancing the details in dark and bright areas. The proposed method successfully restores color fidelity, enhances image contrast and sharpness, and simultaneously preserves details of the original scene. Extensive experiments on the public dataset DAVIS-NUIUIED and our dataset DAVIS-MDUIED demonstrate that the proposed method outperforms state-of-the-art methods in enhancing multi-degraded underwater images.

Traditional trajectory learning methods based on Imitation Learning (IL) only learn the existing trajectory knowledge from human demonstration. In this way, it can not adapt the trajectory knowledge to the task environment by interacting with the environment and fine-tuning the policy. To address this problem, a global trajectory learning method which combinines IL with Reinforcement Learning (RL) to adapt the knowledge policy to the environment is proposed. In this paper, IL is proposed to acquire basic trajectory skills, and then learns the agent will explore and exploit more policy which is applicable to the current environment by RL. The basic trajectory skills include the knowledge policy and the time stage information in the whole task space to help learn the time series of the trajectory, and are used to guide the subsequent RL process. Notably, neural networks are not used to model the action policy and the Q value of RL during the RL process. Instead, they are sampled and updated in the whole task space and then transferred to the networks after the RL process through Behavior Cloning (BC) to get continuous and smooth global trajectory policy. The feasibility and the effectiveness of the method was validated in a custom Gym environment of a flower drawing task. And then, we executed the learned policy in the real-world robot drawing experiment.

IntroductionReinforcement learning has been widely used in robot motion planning. However, for multi-step complex tasks of dual-arm robots, the trajectory planning method based on reinforcement learning still has some problems, such as ample exploration space, long training time, and uncontrollable training process. Based on the dual-agent depth deterministic strategy gradient (DADDPG) algorithm, this study proposes a motion planning framework constrained by the human joint angle, simultaneously realizing the humanization of learning content and learning style. It quickly plans the coordinated trajectory of dual-arm for complex multi-step tasks.MethodsThe proposed framework mainly includes two parts: one is the modeling of human joint angle constraints. The joint angle is calculated from the human arm motion data measured by the inertial measurement unit (IMU) by establishing a human-robot dual-arm kinematic mapping model. Then, the joint angle range constraints are extracted from multiple groups of demonstration data and expressed as inequalities. Second, the segmented reward function is designed. The human joint angle constraint guides the exploratory learning process of the reinforcement learning method in the form of step reward. Therefore, the exploration space is reduced, the training speed is accelerated, and the learning process is controllable to a certain extent.Results and discussionThe effectiveness of the framework was verified in the gym simulation environment of the Baxter robot's reach-grasp-align task. The results show that in this framework, human experience knowledge has a significant impact on the guidance of learning, and this method can more quickly plan the coordinated trajectory of dual-arm for multi-step tasks.

Background: Robot-assisted fracture reduction systems can potentially reduce the risk of infection and improve outcomes, leading to significant health and economic benefits. However, these systems are still in the laboratory stage and not yet ready for commercialization due to unresolved difficulties. While previous reviews have focused on individual technologies, system composition, and surgical stages, a comprehensive review is necessary to assist future scholars in selecting appropriate research directions for clinical use.Methods: A literature review using Google Scholar identified articles on robot-assisted fracture reduction systems. A comprehensive search yielded 17,800, 18,100, and 16,700 results for “fracture reduction,” “computer-assisted orthopedic surgery,” and “robot-assisted fracture reduction,” respectively. Approximately 340 articles were selected, and 90 highly relevant articles were chosen for further reading after reviewing the abstracts.Results and Conclusion: Robot-assisted fracture reduction systems offer several benefits, including improved reduction accuracy, reduced physical work and radiation exposure, enhanced preoperative planning and intraoperative visualization, and shortened learning curve for skill acquisition. In the future, these systems will become integrated and practical, with automatic preoperative planning and high intraoperative safety.

IntroductionBehavioral Cloning (BC) is a common imitation learning method which utilizes neural networks to approximate the demonstration action samples for task manipulation skill learning. However, in the real world, the demonstration trajectories from human are often sparse and imperfect, which makes it challenging to comprehensively learn directly from the demonstration action samples. Therefore, in this paper, we proposes a streamlined imitation learning method under the terse geometric representation to take good advantage of the demonstration data, and then realize the manipulation skill learning of assembly tasks.MethodsWe map the demonstration trajectories into the geometric feature space. Then we align the demonstration trajectories by Dynamic Time Warping (DTW) method to get the unified data sequence so we can segment them into several time stages. The Probability Movement Primitives (ProMPs) of the demonstration trajectories are then extracted, so we can generate a lot of task trajectories to be the global strategy action samples for training the neural networks. Notalby, we regard the current state of the assembly task as the via point of the ProMPs model to get the generated trajectories, while the time point of the via point is calculated according to the probability model of the different time stages. And we get the action of the current state according to the target position of the next time state. Finally, we train the neural network to obtain the global assembly strategy by Behavioral Cloning.ResultsWe applied the proposed method to the peg-in-hole assembly task in the simulation environment based on Pybullet + Gym to test its task skill learning performance. And the learned assembly strategy was also executed on a real robotic platform to verify the feasibility of the method further.DiscussionAccording to the result of the experiment, the proposed method achieves higher success rates compared to traditional imitation learning methods while exhibiting reasonable generalization capabilities. It shows that the ProMPs under geometric representation can help the BC method make better use of the demonstration trajectory and thus better learn the task skills.

In this paper, we propose a deep reinforcement learning-based framework that enables adaptive and continuous control of a robot to push unseen objects from random positions to the target position. Our approach takes into account contact information in the design of the reward function, resulting in improved success rates, generalization for unseen objects, and task efficiency compared to policies that do not consider contact information. Through reinforcement learning using only one object in simulation, we obtain a learned policy for manipulating a single object, which demonstrates good generalization when applied to the task of pushing unseen objects. Finally, we validate the effectiveness of our approach in real-world scenarios.

Accurate recognition of patients’ movement intentions and real-time adjustments are crucial in rehabilitation exoskeleton robots. However, some patients are unable to utilize electromyography (EMG) signals for this purpose due to poor or missing signals in their lower limbs. In order to address this issue, we propose a novel method that fits gait parameters using cerebral blood oxygen signals. Two types of walking experiments were conducted to collect brain blood oxygen signals and gait parameters from volunteers. Time domain, frequency domain, and spatial domain features were extracted from brain hemoglobin. The AutoEncoder-Decoder method is used for feature dimension reduction. A regression model based on the long short-term memory (LSTM) model was established to fit the gait parameters and perform incremental learning for new individual data. Cross-validation was performed on the model to enhance individual adaptivity and reduce the need for individual pre-training. The coefficient of determination (R2) for the gait parameter fit was 71.544%, with a mean square error (RMSE) of less than 3.321%. Following adaptive enhancement, the coefficient of R2 increased by 6.985%, while the RMSE decreased by 0.303%. These preliminary results indicate the feasibility of fitting gait parameters using cerebral blood oxygen information. Our research offers a new perspective on assisted locomotion control for patients who lack effective myoelectricity, thereby expanding the clinical application of rehabilitation exoskeleton robots. This work establishes a foundation for promoting the application of Brain-Computer Interface (BCI) technology in the field of sports rehabilitation.

The use of robots in various industries is evolving from mechanization to intelligence and precision. These systems often comprise parts made of different materials and thus require accurate and comprehensive target identification. While humans perceive the world through a highly diverse perceptual system and can rapidly identify deformable objects through vision and touch to prevent slipping or excessive deformation during grasping, robot recognition technology mainly relies on visual sensors, which lack critical information such as object material, leading to incomplete cognition. Therefore, multimodal information fusion is believed to be key to the development of robot recognition. Firstly, a method of converting tactile sequences to images is proposed to deal with the obstacles of information exchange between different modalities for vision and touch, which overcomes the problems of the noise and instability of tactile data. Subsequently, a visual-tactile fusion network framework based on an adaptive dropout algorithm is constructed, together with an optimal joint mechanism between visual information and tactile information established, to solve the problem of mutual exclusion or unbalanced fusion in traditional fusion methods. Finally, experiments show that the proposed method effectively improves robot recognition ability, and the classification accuracy is as high as 99.3%.

ObjectivesThe objective of this study was to investigate common functional near-infrared spectroscopy (fNIRS) features of mental fatigue induced by different tasks. In addition to distinguishing fatigue from non-fatigue state, the early signs of fatigue were also studied so as to give an early warning of fatigue.MethodsfNIRS data from 36 participants were used to investigate the common character of functional connectivity network corresponding to mental fatigue, which was induced by psychomotor vigilance test (PVT), cognitive work, or simulated driving. To analyze the network reorganizations quantitatively, clustering coefficient, characteristic path length, and small worldness were calculated in five sub-bands (0.6–2.0, 0.145–0.600, 0.052–0.145, 0.021–0.052, and 0.005–0.021 Hz). Moreover, we applied a random forest method to classify three fatigue states.ResultsIn a moderate fatigue state: the functional connectivity strength between brain regions increased overall in 0.021–0.052 Hz, and an asymmetrical pattern of connectivity (right hemisphere > left hemisphere) was presented. In 0.052–0.145 Hz, the connectivity strength decreased overall, the clustering coefficient decreased, and the characteristic path length increased significantly. In severe fatigue state: in 0.021–0.052 Hz, the brain network began to deviate from a small-world pattern. The classification accuracy of fatigue and non-fatigue was 85.4%. The classification accuracy of moderate fatigue and severe fatigue was 82.8%.ConclusionThe preliminary research demonstrates the feasibility of detecting mental fatigue induced by different tasks, by applying the functional network features of cerebral hemoglobin signal. This universal and robust method has the potential to detect early signs of mental fatigue and prevent relative human error in various working environments.

The hippocampus and its accessory are the main areas for spatial cognition. It can integrate paths and form environmental cognition based on motion information and then realize positioning and navigation. Learning from the hippocampus mechanism is a crucial way forward for research in robot perception, so it is crucial to building a calculation method that conforms to the biological principle. In addition, it should be easy to implement on a robot. This paper proposes a bionic cognition model and method for mobile robots, which can realize precise path integration and cognition of space. Our research can provide the basis for the cognition of the environment and autonomous navigation for bionic robots.

Past research indicates that childhood emotional maltreatment (CEM) undermines the quality of adult romantic relationships by fostering negative characteristics in survivors. Two longitudinal studies investigated the hypothesis that decreased compassionate goals toward partners over time explain the association between CEM and declining relationship quality. In Study 1, CEM predicted decreased compassionate goals over time, which in turn predicted decreased relationship quality in individuals in romantic relationships. Study 2 replicated this effect in romantically involved couples and showed that partners’ high compassionate goals attenuated the decline in compassionate goals associated with reported CEM. These results point to the importance of examining how CEM may affect positive relationship processes and the protective roles of partners’ compassionate goals.

Many algorithms in probabilistic sampling-based motion planning have been proposed to create a path for a robot in an environment with obstacles. Due to the randomness of sampling, they can efficiently compute the collision-free paths made of segments lying in the configuration space with probabilistic completeness. However, this property also makes the trajectories have some unnecessary redundant or jerky motions, which need to be optimized. For most robotics applications, the trajectories should be short, smooth and keep away from obstacles. This paper proposes a new trajectory optimization technique which transforms a polygon collision-free path into a smooth path, and can deal with trajectories which contain various task constraints. The technique removes redundant motions by quadratic programming in the parameter space of trajectory, and converts collision avoidance conditions to linear constraints to ensure absolute safety of trajectories. Furthermore, the technique uses a projection operator to realize the optimization of trajectories which are subject to some hard kinematic constraints, like keeping a glass of water upright or coordinating operation with dual robots. The experimental results proved the feasibility and effectiveness of the proposed method, when it is compared with other trajectory optimization methods.

ObjectiveTo investigate the usefulness of positron emission tomography (PET) images obtained after carbon-ion irradiation for dose verification in carbon-ion radiotherapy.Methods and MaterialsAn anthropomorphic head phantom was used in this study. Three cubes with volumes of 1, 4, and 10 ml were contoured as targets in the phantom CT through a treatment planning system. Treatment plans with six prescriptions from 2.5 to 10 Gy (2.5, 3, 5, 6, 8, and 10 Gy effective dose) were designed and delivered by 90° fixed carbon-ion beams, respectively. After irradiation of the phantom, a PET/CT scan was performed to fuse the treatment-planning CT image with the PET/CT image. The relationship between target volume and the standard uptake value (SUV) in PET/CT was evaluated for corresponding plan prescription. The MIM Maestro software was used for the image fusion and data analysis.ResultsSUV in the target had an approximate linear relationship with the effective dose. The same effective dose could generate a roughly equal SUV for different target volumes (p < 0.05).ConclusionsIt is feasible to verify the actual 3-D dose distribution of carbon-ion radiotherapy by the approach in this study.

With the developing need for luminous materials with better performance, lanthanide-doped nanocrystals have been widely studied for their unique luminescence properties such as their narrow bandwidth emission, excellent chemical stability, and photostability, adjustable emission color, high signal-to-background ratio, deeper tissue penetration with less photo-damage, and low toxicity, etc., which triggered enthusiasm for research on the broad applications of lanthanide-doped nanocrystals in bioimaging, anti-counterfeiting, biosensing, and cancer diagnosis and treatment. Considerable progress has been made in the past few decades, but low upconversion luminescence efficiency has been a hindrance in achieving further progress. It is necessary to summarize the recently relevant literature and find solutions to improve the efficiency. The latest experimental and theoretical studies related to the deliberate design of rare earth luminescent nanocrystals have, however, shown the development of metal ion-doped approaches to enhance the luminescent intensity. Host lattice manipulation can enhance the luminescence through increasing the asymmetry, which improves the probability of electric dipole transition; and the energy transfer modulation offers a reduced cross-relaxation pathway to improve the efficiency of the energy transfer. Based on the mechanisms of host lattice manipulation and energy transfer modulation, a wide range of enhancements at all wavelengths or even within a particular wavelength have been accomplished with an enhancement of up to a hundred times. In this mini review, we present the strategy of metal ion-doped lanthanide nanocrystals to cope with the issue of enhancing luminescence, overview the advantages and tricky challenges in boosting the luminescence, and provide a potential trend of future study in this field.

In robotic radiosurgery, motion tracking is crucial for accurate treatment planning of tumor inside the thoracic or abdominal cavity. Currently, motion characterization for respiration tracking mainly focuses on markers that are placed on the surface of human chest. Nevertheless, limited markers are not capable of expressing the comprehensive motion feature of the human chest and abdomen. In this paper, we proposed a method of respiratory motion characterization based on the voxel modeling of the thoracoabdominal torso. Point cloud data from depth cameras were used to achieve three-dimensional modeling of the chest and abdomen surface during respiration, and a dimensionality reduction algorithm was proposed to extract respiratory features from the established voxel model. Finally, experimental results including the accuracy of voxel model and correlation coefficient were compared to validate the feasibility of the proposed method, which provides enhanced accuracy of target motion correlation than traditional methods that utilized external markers.

This paper describes an improved brain-inspired simultaneous localization and mapping (RatSLAM) that extracts visual features from saliency maps using a frequency-tuned (FT) model. In the traditional RatSLAM algorithm, the visual template feature is organized as a one-dimensional vector whose values only depend on pixel intensity; therefore, this feature is susceptible to changes in illumination intensity. In contrast to this approach, which directly generates visual templates from raw RGB images, we propose an FT model that converts RGB images into saliency maps to obtain visual templates. The visual templates extracted from the saliency maps contain more of the feature information contained within the original images. Our experimental results demonstrate that the accuracy of loop closure detection was improved, as measured by the number of loop closures detected by our method compared with the traditional RatSLAM system. We additionally verified that the proposed FT model-based visual templates improve the robustness of familiar visual scene identification by RatSLAM.

Lanthanide-doped upconversion nanoparticles can convert long wavelength excitation radiation to short wavelength emission. They have great potential in biomedical applications, such as bioimaging, biodetection, drug delivery, and theranostics. However, there is little information available on their bioavailability and biological effects after oral administration. In this study, we systematically investigated the bioavailability, biodistribution, and toxicity of silica-coated upconversion nanoparticles administrated by gavage. Our results demonstrate that these nanoparticles can permeate intestinal barrier and enter blood circulation by microstructure observation of Peyer's patch in the intestine. Comparing the bioavailability and the biodistribution of silica-coated upconversion nanoparticles with oral and intravenous administration routes, we found that the bioavailability and biodistribution are particularly dependent on the administration routes. After consecutive gavage for 14 days, the body weight, pathology, Zn and Cu level, serum biochemical analysis, oxidative stress, and inflammatory cytokines were studied to further evaluate the potential toxicity of the silica-coated upconversion nanoparticles. The results suggest that these nanoparticles do not show overt toxicity in mice even at a high dose of 100 mg/kg body weight.

People often take either the role of an actor or that of recipient in positive and negative interpersonal events when they interact with others. The present study investigated how the actor-recipient role affected the neural responses to self in emotional situations. Twenty-five participants were scanned while they were presented with positive and negative interpersonal events and were asked to rate the degree to which the actor/the recipient was that kind of person who caused the interpersonal event. Half of the trials were self-relevant events and the other half were other-relevant events. Results showed that people were more likely to isolate self from negative events when they played the role of actor relative to recipient. Pregenual anterior cingulate cortex (pgACC) and posterior dorsal anterior cingulate cortex (pdACC) were more active for self than other only in negative events. More importantly, also in negative interpersonal events, dorsal medial prefrontal cortex (dmPFC) showed greater self-related activations (self-other) when participants played the role of recipient relative to actor, while activities in orbitofrontal cortex (OFC) were greater for self than other only when the evaluation target played the role of recipient. These results showed that the actor-recipient role affected neural responses to self in emotional situations, especially when a recipient role was played in negative situations. [ABSTRACT FROM AUTHOR]