Contexts play an important role in salient object detection (SOD). High-level contexts describe the relations between different parts/objects and thus are helpful for discovering the specific locations of salient objects while low-level contexts could provide the fine detail information for delineating the boundary of the salient objects. However, the way of perceiving/leveraging rich contexts has not been fully investigated by existing SOD works. The common context extraction strategies ( e.g. , leveraging convolutions with large kernels or atrous convolutions with large dilation rates) do not consider the effectiveness and efficiency simultaneously and may cause sub-optimal solutions. In this paper, we devote to exploring an effective and efficient way to learn rich contexts for accurate SOD. Specifically, we first build a dense context exploration (DCE) module to capture dense multi-scale contexts and further leverage the learned contexts to enhance the features discriminability. Then, we embed multiple DCE modules in an encoder-decoder architecture to harvest dense contexts of different levels. Furthermore, we propose an attentive skip-connection to transmit useful features from the encoder part to the decoder part for better dense context exploration. Finally, extensive experiments demonstrate that the proposed method achieves more superior detection results on the six benchmark datasets than 18 state-of-the-art SOD methods.

Human motion prediction intends to predict how humans move given a historical sequence of 3D human motions. Recent transformer-based methods have attracted increasing attentions and demonstrated their promising performance in 3D human motion prediction. However, existing methods generally decompose the input of human motion information into spatial and temporal branches in a separate way and seldom consider their inherent coherence between the two branches, hence often failing to register the dynamic spatio-temporal information during the training process. Motivated by these issues, we propose a spatio-temporal cross-transformer network (STCT) for 3D human motion predictions. Specifically, we investigate various types of interaction methods (i.e., Concatenation Interaction, Msg token interaction, and Cross-transformer) to capture the coherence of the spatial and temporal branches. According to the obtained results, the proposed cross-transformer interaction method shows its superiority over other methods. Meanwhile, considering that most existing works treat the human body as a set of 3D human joint positions, the predicted human joints are proportionally less appropriate to the realistic human body due to unreasonable bone length and non-plausible poses as time progresses. We further resort to the bone constraints of human mesh to produce more realistic human motions. By fitting a parametric body model (i.e., SMPL-X model) to the predicted human joints, a reconstruction loss function is proposed to remedy the unreasonable bone length and pose errors. Comprehensive experiments on AMASS and Human3.6M datasets have demonstrated that our method achieves superior performance over compared methods.

The emergence of the right to be forgotten has sparked interest in federated unlearning. Researchers utilize federated unlearning to address the issue of removing user contributions. However, practical implementation is challenging when malicious clients are involved. In this paper, we propose FAST, i.e., Adopting F ederated unle A rning to Eliminating Maliciou S T erminals at Server Side, a framework with three main components: 1) eliminating contributions of malicious clients: the central server records and subtracts updates of malicious clients from the global model, 2) judging unlearning efficiency: we model a mechanism to assess unlearning efficiency and prevent over-unlearning, and 3) remedying unlearning model performance: the central server utilizes a benchmark dataset to remedy model bias during unlearning. Experimental results demonstrate that FAST can achieve 96.98% accuracy on the MNIST dataset with 40% malicious clients, offering a 16x speedup to retraining from scratch. Meanwhile, it can recover model utility with high efficiency, and extensive evaluations of four real-world datasets demonstrate the validity of our proposed scheme.

Current mainstream pedestrian detectors tend to profit directly from convolutional neural networks (CNNs) that are designed for image classification. While requiring a large downsampling factor to produce high-level semantic features, CNNs cannot adaptively focus on the useful channels and regions of the feature maps, which limits the accuracy of pedestrian detection. To obtain a higher accuracy, we propose a single-stage pedestrian detector with channel and spatial attentions (CSANet), which can locate useful channels and regions automatically while extracting features. The backbone of CSANet is different from that of mainstream pedestrian detectors, which can effectively highlight the pedestrian-likely regions and suppress the background. Specifically, we model contextual dependencies from channel and spatial dimensions of the feature maps, respectively. The channel attention module can selectively promote CNNs to focus on key channels by integrating associated features. Meantime, the spatial attention module can illuminate semantic pixels by aggregating similar features of all channels. Eventually, the two modules are connected in series to further enhance the representation of feature maps. Experiment results show that CSANet achieves the state-of-the-art performance with MR-2 of 3.55% on Caltech dataset and obtains competitive performance on CityPersons dataset while maintaining a high computational efficiency.

With the development of aerospace engineering, the space on-orbit servicing has been brought more attention to many scholars. Obstacle avoidance planning of space manipulator end-effector also attracts increasing attention. This problem is complex due to the existence of obstacles. Therefore, it is essential to avoid obstacles in order to improve planning of space manipulator end-effector. In this paper, we proposed an improved ant colony algorithm to solve this problem, which is effective and simple. Firstly, the models were established respectively, including the kinematic model of space manipulator and expression of valid path in space environment. Secondly, we described an improved ant colony algorithm in detail, which can avoid trapping into local optimum. The search strategy, transfer rules, and pheromone update methods were all adjusted. Finally, the improved ant colony algorithm was compared with the classic ant colony algorithm through the experiments. The simulation results verify the correctness and effectiveness of the proposed algorithm.