Photorealistic frontal view synthesis from a single face image has wide range of applications in the field recognition. Although data-driven deep learning methods have been proposed to address this problem by seeking solutions ample data, is still challenging because it intrinsically ill-posed. This paper proposes Two-Pathway Generative Adversarial Network (TP-GAN) for photorealistic simultaneously perceiving global structures and local details. Four landmark located patch networks are...

Spotting objects that are visually adapted to their surroundings is challenging for both humans and AI. Conventional generic / salient object detection techniques suboptimal this task because they tend only discover easy clear objects, while overlooking the difficult-to-detect ones with inherent uncertainties derived from indistinguishable textures. In work, we contribute a novel approach using probabilistic representational model in combination transformers explicitly reason under...

Human-powered lower exoskeletons have received considerable interests from both academia and industry over the past decades, encountered increasing applications in human locomotion assistance strength augmentation. One of most important aspects those is to achieve robust control exoskeletons, which, first place, requires proactive modeling movement trajectories through physical human-robot interaction (pHRI). As a powerful representative tool for motion trajectories, dynamic primitives (DMP)...

Photorealistic frontal view synthesis from a single face image has wide range of applications in the field recognition. Although data-driven deep learning methods have been proposed to address this problem by seeking solutions ample data, is still challenging because it intrinsically ill-posed. This paper proposes Two-Pathway Generative Adversarial Network (TP-GAN) for photorealistic simultaneously perceiving global structures and local details. Four landmark located patch networks are...

Rapid object recognition in the industrial field is key to intelligent manufacturing. The research on fast methods based deep learning was focus of researchers recent years, but balance between detection speed and accuracy not well solved. In this paper, a method for electronic components complex background presented. Firstly, we built image dataset, including acquisition, augmentation, labeling. Secondly, proposed. solved through lightweight improvement YOLO (You Only Look Once)-V3 network...

Machine vision is one of the key technologies used to perform intelligent manufacturing. In order improve recognition rate multi-class defects in wheel hubs, an improved Faster R-CNN method was proposed. A data set for hub built. This consisted four types 2,412 1080 × 1440 pixels images. modified, trained, verified and tested based on this database. The proposed excellent. compared with popular YOLOv3 methods showing simpler, faster, more accurate defect detection, which demonstrates...

To deal with plant−model mismatches in control practice, this paper proposes two variations of an offset-free framework which integrates nonlinear model predictive (NMPC) and moving horizon estimation (MHE). We prove that the proposed method achieves regulatory behavior, even presence mismatches. If plant uncertainty structure is known, MHE can be tuned to estimate parameters, remove mismatch online. In addition, we incorporate advanced step NMPC (as-NMPC) (as-MHE) strategies into reduce...

Learning by demonstration methods have gained considerable interest in human-coupled robot control. It aims at modeling the goal motion trajectories through human demonstration. However, lower exoskeleton control, physical human-robot interaction is changing from pilot to or even for one different walking patterns. This characteristic requires that exoskeletons should ability learn and adapt as well controllers online. paper presents a novel Hierarchical Interactive (HIL) strategy which...

To investigate an optimal strategy by assessing the effectiveness of varying follicular sizes on trigger day during luteal phase stimulation protocol and provide evidence for personalized adjustment. This was a retrospective study including total 661 patients who had started their in vitro fertilization cycle with (LPS) 2015–2023. We classified into groups according to size dominant proportion follicles human chorionic gonadotropin (hCG) day: large, medium, small. The metaphase II (MII)...

End-to-end deep learning has gained considerable interests in autonomous driving vehicles both academic and industrial fields, especially decision making process. One critical issue process of is steering control. Researchers already trained different artificial neural networks to predict angle with front-facing camera data stream. However, existing end-to-end methods only consider the spatiotemporal relation on a single layer lack ability extracting future information. In this paper, we...

Purpose . Powered lower-limb exoskeleton has gained considerable interests, since it can help patients with spinal cord injury(SCI) to stand and walk again. Providing walking assistance SCI patients, most exoskeletons are designed follow predefined gait trajectories, which makes the patient unnaturally feels uncomfortable. Furthermore, trajectories cannot always maintain balance especially when encountering disturbances. Design/Methodology/Approach This paper proposed a novel planning...

Sensitivity Amplification Control (SAC) algorithm was first proposed in the augmentation applications of Berkeley Lower Extremity Exoskeleton (BLEEX). The SAC is widely used human since it just need information from exoskeleton robot, so that complexity system can be reduced greatly. However, has two main drawbacks: 1) requiring accurate dynamic models exoskeleton, 2) not manage variation interaction dynamics different walking speed. This paper presents a novel developed learning control...

Learning skills autonomously is a particularly important ability for an autonomous robot. A promising approach reinforcement learning (RL) where agents learn policy through interaction with its environment. One problem of RL algorithm how to tradeoff the exploration and exploitation. Moreover, multiple tasks also make great challenge robot learning. In this paper, enhance performance RL, novel framework integrating knowledge transfer proposed. Three basic components are included: 1)...

Lower limb exoskeleton (LLE) has received considerable interests in strength augmentation, rehabilitation and walking assistance scenarios. For assistance, the LLE is expected to have capability of controlling affected leg track unaffected leg's motion naturally. An important issue this scenario that system needs deal with unpredictable disturbance from patient, which requires controller ability adapt different wearers. This paper proposes a novel Data-Driven Reinforcement Learning (DDRL)...

In this brief, a tracking control problem for robotic systems with unknown uncertainties is addressed by using an event-triggered adaptive dynamic programming (ADP) method. First, the of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$n$ </tex-math></inline-formula> -degree freedom (DOF) system transformed to optimal auxiliary such that robust design original feasible based on ADP framework. To reduce...