Head pose estimation plays a crucial role in various applications, including human-machine interaction, autonomous driving systems, and3d reconstruction. Current methods address the problem primarily from a 2d perspe...
详细信息
Head pose estimation plays a crucial role in various applications, including human-machine interaction, autonomous driving systems, and3d reconstruction. Current methods address the problem primarily from a 2d perspective, which limits the efficient utilization of 3d information. Herein, a novel approach, called pose orientation-aware network (POANet), which leverages normal maps for orientation information embedding, providing abundant and robust head pose information, is introduced. POANet incorporates the axial signal perception module and the rotation matrix perception module, these lightweight modules make the approach achieve state-of-the-art (SOTA) performance with few computational costs. This method can directly analyze various topological 3ddata without extensive preprocessing. For depth images, POANet outperforms existing methods on the Biwi Kinect head pose dataset, reducing the mean absolute error (MAE) by approximate to 30% compared to the SOTA methods. POANet is the first method to perform rigid head registration in a landmark-free manner. It also incorporates few-shot learning capabilities and achieves an MAE of about 1 degrees$1<^>{\circ}$ on the Headspace dataset. These features make POANet a superior alternative to traditional generalized Procrustes analysis for mesh data processing, offering enhanced convenience for human phenotype studies. Pose orientation-aware network (POANet), a lightweight model for 3d head pose estimation that analyzes diverse topological 3ddata, is introduced. POANet outperforms state-of-the-art method on the BIWI depth imagedataset, reducing the mean absolute error by approximate to 30%. Additionally, it is the first solution to perform rigid head registration in a landmark-free manner on mesh *** (c) 2024 WILEY-VCH GmbH
In this article, we propose a generative adversarial network based fringe pattern normalization method. We investigate the method’s effectiveness under various noise levels by evaluating root mean square error (RMSE)...
详细信息
We introduce light-field microscopy for volumetric imaging of 3d Human induced pluripotent stem cell-derived cardiomyocytes with high spatiotemporal resolution, realizing simultaneous recording of calcium oscillations...
详细信息
In this paper, homogeneous as well as spatially varying polarization distributions of Bessel beams is studied using spiral axicon hologram displayed on spatial light modulator. Theoretical and experimental results for...
详细信息
We present the Fourier Light field Camera Array Microscope (FL-CAM) for high-throughput, single-snapshot 3d imaging. The FL-CAM substitutes a synchronized array of 48 independent imaging systems for micro-lens array o...
详细信息
We propose a neural network regularizer that mitigates the ill-conditionedness of the Lippmann-Schwinger equation. By learning the physics, it can significantly reduce the computational time and can be generalizable t...
详细信息
Angular multiplexing polarization holography was proposed to realize the measurement of the complete Stokes-Mueller matrix through only 6 recordings. Measurement experiments of isotropic and anisotropic media were con...
详细信息
Since single sensor mask may cause information lost during phase retrieval, we refer to the Monte Carlo method to design multiple binary sensor masks for one diffraction pattern, leading to the enhanced phase retrieva...
详细信息
We propose a photonic neural network and its experimental demonstration that enables high-throughput and low-latency image transmission by implementing most processing in optical domain, reducing latency of computatio...
详细信息
暂无评论