To improve the accuracy of cross-domain object detection, the existing unsupervised domain adaptation (UDA) object detection methods mostly use Feature Pyramid Network (FPN), multiple Region Proposal Network (RPN), and multiple domain classifier, but these methods lead to complex network structures, slow model convergence, and low detection efficiency. To solve the above problems, this paper proposes an Efficient and Accurate Cross-domain Object Detection Method Using One-level Feature and Domain Adaptation (OFDA). This method realizes one-level feature object detection through feature fusion and divide-and-conquer technology; realizes overfitting feature suppression and unsupervised domain adaptation through domain-specific suppression and domain feature alignment technology; realizes background feature suppression through the Objectness branch, which replaces the time-consuming Region Proposal Network (RPN) structure and improves the efficiency of unsupervised adaptive detection. The paper verifies the feasibility and superiority of the proposed method by the comparative experiments and ablation experiments of multiple datasets. The proposed OFDA method not only improves the efficiency of object detection, but also ensures the accuracy of cross-domain detection.
Abstract Background: Among the complications related to colostomy, variceal hemorrhage caused by cirrhosis is a rare case. The occurrence of peristome variceal hemorrhage should be rapid and effective treatment, otherwise it will result in hemorrhagic shock, repeated stoma bleeding, systemic organs deterioration and other adverse events. Case presentation: In this case of liver cirrhosis, Stoma varices and repeated bleeding due to decompensated cirrhosis was performed by local compression and suture therapy but failed with rebleeding. Being combined with abdominal enhanced CT images，Stoma varices are derived from the branch of inferior mesenteric vein, whose above trunk is located and highly ligated and vein balls are stripped. Meanwhile, the Stoma below abdominal wall varicose veins is sutured by intracutaneous way, which could achieve satisfactory hemostasis in intra-operation. Abdominal enhanced CT images prompted that peristome varicose veins are disappeared. The skin around the stoma is flat and the mucosa is ruddy when changing the bag in post-operation. Conclusion: The report that our team reduced the risk of peristomal variceal bleeding by blocking the main variceal vein through a small incision under local anesthesia may open up another treatment strategy for such patients.
The separation and reconfiguration of stacked satellites in orbit is an effective technique for constructing large space structures. The natural coordinates formulation is used to establish the dynamic equations for stacked satellite systems, which has the advantage of facilitating the handling of fixed constraints between satellites. Suitable strategies for autonomous assembly separation and assembly are devised. A spin separation method is employed to achieve collision-free separation of satellites, while PD control and the potential function is utilized for satellite assembly. Additionally, an optimization algorithm is employed to calculate the minimum distance between satellites, enabling precise determination of the potential function's magnitude. By implementing these methods in simulations, the complete process from separation of stacked satellites to segmented assembly is realized, which confirming the effectiveness of the proposed separation and assembly strategies.
Physical adversarial attacks in object detection have attracted increasing attention. However, most previous works focus on hiding the objects from the detector by generating an individual adversarial patch, which only covers the planar part of the vehicle's surface and fails to attack the detector in physical scenarios for multi-view, long-distance and partially occluded objects. To bridge the gap between digital attacks and physical attacks, we exploit the full 3D vehicle surface to propose a robust Full-coverage Camouflage Attack (FCA) to fool detectors. Specifically, we first try rendering the nonplanar camouflage texture over the full vehicle surface. To mimic the real-world environment conditions, we then introduce a transformation function to transfer the rendered camouflaged vehicle into a photo realistic scenario. Finally, we design an efficient loss function to optimize the camouflage texture. Experiments show that the full-coverage camouflage attack can not only outperform state-of-the-art methods under various test cases but also generalize to different environments, vehicles, and object detectors. The code of FCA will be available at: https://idrl-lab.github.io/Full-coverage-camouflage-adversarial-attack/.
The paper focuses on the separation and deployment dynamics of an on-orbit compactly connected multi-rigid-body (MRB) system, which could separate autonomously from a carrier spacecraft. Based on the focused MRB system, it is not necessary to repeatedly use the launcher of the carrier spacecraft or install multiple launchers in the spacecraft to separate the MRB system. This is advantageous because it can effectively improve the space utilization rate of the spacecraft, simplify the separation deployment operations and reduce the risk of collision between rigid bodies. To realize the separation of such a MRB system, the paper presents an investigation on its on-orbit dynamics and the design of collision-free separation deployment schemes. Firstly, a dynamic model of a single rigid body is established based on the principle of virtual work and the Natural Coordinate Formulation (NCF) method accounting for the relative motion between rigid bodies and attitude changes of each rigid body. Considering the orbital motion, the variations of connecting constraints of the MRB system and the interactions between rigid bodies during the separation, the governing nonlinear dynamic equations including constraints of the system are obtained with a method of Lagrange multipliers. With practical engineering applications taken into consideration, the separation deployment of MRB system is realized through ejection mechanisms mounted on the four corners of each contact surface between rigid bodies. Secondly, the timing sequences of separation maneuvers are specially programmed and two separation schemes are developed by adjusting different ejection directions and ejection sequences to guarantee the non-collision between rigid bodies in the separation deployment. Finally, numerical case studies are presented for investigating the nonlinear dynamic behaviors of rigid bodies and demonstrating the effectiveness of separation schemes.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)