Search

In this work, we propose a novel method for head pose estimation and face recovery, particularly to solve the potential impacts of noises in signal processing to get an efficient and effective model that is more resilient with annoying effects through adding affine transformation with the low-rank robust subspace regression. Consequently, the corrupted images can be correctly recovered by affine transformations to render more best regression outcomes. Thereby, we need to search so as to get optimal parameters which can be regarded as convex constrained optimization techniques. Afterward, the alternating direction method for multipliers (ADMM) approach is considered and a new set of updated equations is well established so as to update the optimization parameters and affine transformations iteratively in a round-robin manner. Additionally, the convergence of these new updating equations is well scrutinized as well. Thus, the experimental simulations reveal that the proposed method outperforms the state-of-the-art works for head pose estimation and face recovery on some public databases. [ABSTRACT FROM AUTHOR]

In order to improve industrial production efficiency, a hand–eye system based on 3D vision is proposed and the proposed system is applied to the assembly task of workpieces. First, a hand–eye calibration optimization algorithm based on data filtering is proposed in this paper. This method ensures the accuracy required for hand–eye calibration by filtering out part of the improper data. Furthermore, the improved U-net is adopted for image segmentation and SAC-IA coarse registration ICP fine registration method is adopted for point cloud registration. This method ensures that the 6D pose estimation of the object is more accurate. Through the hand–eye calibration method based on data filtering, the average error of hand–eye calibration is reduced by 0.42 mm to 0.08 mm. Compared with other models, the improved U-net proposed in this paper has higher accuracy for depth image segmentation, and the Acc coefficient and Dice coefficient achieve 0.961 and 0.876, respectively. The average translation error, average rotation error and average time-consuming of the object recognition and pose estimation methods proposed in this paper are 1.19 mm, 1.27°, and 7.5 s, respectively. The experimental results show that the proposed system in this paper can complete high-precision assembly tasks. [ABSTRACT FROM AUTHOR]