Qixiang Ma

ORCID: 0009-0004-0424-0308
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Computer Graphics and Visualization Techniques
  • Advanced Vision and Imaging
  • 3D Shape Modeling and Analysis
  • Remote Sensing and Land Use
  • Advanced Computational Techniques and Applications
  • Face and Expression Recognition
  • Robotics and Sensor-Based Localization
  • Video Surveillance and Tracking Methods
  • Regional Economic and Spatial Analysis
  • Image Processing and 3D Reconstruction
  • Domain Adaptation and Few-Shot Learning
  • Text and Document Classification Technologies
  • Virtual Reality Applications and Impacts
  • Augmented Reality Applications
  • Interactive and Immersive Displays

Beihang University
2021-2025

Renmin University of China
2008

Rendering quality and performance greatly affect the user's immersion in VR experiences. 3D Gaussian Splatting-based methods can achieve photo-realistic rendering with speeds of over 100 fps static scenes, but speed drops below 10 monocular dynamic scenes. Foveated provides a possible solution to accelerate without compromising visual perceptual quality. However, 3DGS foveated are not compatible. In this paper, we propose Fov-GS, splatting method for scenes real time. We introduce forest...

10.1109/tvcg.2025.3549576 article EN IEEE Transactions on Visualization and Computer Graphics 2025-01-01

We introduce the disocclusion headlight, a method for VR selection assistance based on alleviating occlusions at center of user's field view. The visualization VE is modified to reduce overlap between objects. This way, candidate objects have larger image footprints, which facilitates selection. modification confined frame, with continuity periphery frame rendered conventionally. provided automatically, without any interaction from user. Furthermore, our disoccludes destroying local spatial...

10.1109/vr50410.2021.00043 article EN 2021-03-01

We propose a new scene-aware foveated rendering method, which incorporates the scene awareness and characteristics of human visual system into mapping-based framework. First, we generate conservative importance map that encodes features scene, acuity, gaze motion. Second, construct pixel size control using convolution kernel method. Third, utilize to guide rendering. At last, temporal coherent refinement strategy is used maintain smooth for adjacent frames. Compared state-of-the-art methods...

10.1109/tvcg.2024.3456157 article EN IEEE Transactions on Visualization and Computer Graphics 2024-01-01

An integrated approach of ISODATA and SVR is presented to extract the objective information, e.g. wheat, which can adequately combine advantages both hard soft classification. It exploits classification method for typical feature mixed spectral unmixing feature. The accuracy assessment shows that this method, obtain a higher than either linear or practical.

10.1117/12.813212 article EN Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE 2008-10-23
Coming Soon ...