Search

Your search keyword '"Shi, Shengxian"' showing total 127 results
Start Over Author "Shi, Shengxian"
127 results on '"Shi, Shengxian"'

9. VommaNet: an End-to-End Network for Disparity Estimation from Reflective and Texture-less Light Field Images

Author

Ma, Haoxin, Li, Haotian, Qian, Zhiwen, Shi, Shengxian, and Mu, Tingting

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

The precise combination of image sensor and micro-lens array enables lenslet light field cameras to record both angular and spatial information of incoming light, therefore, one can calculate disparity and depth from light field images. In turn, 3D models of the recorded objects can be recovered, which is a great advantage over other imaging system. However, reflective and texture-less areas in light field images have complicated conditions, making it hard to correctly calculate disparity with existing algorithms. To tackle this problem, we introduce a novel end-to-end network VommaNet to retrieve multi-scale features from reflective and texture-less regions for accurate disparity estimation. Meanwhile, our network has achieved similar or better performance in other regions for both synthetic light field images and real-world data compared to the state-of-the-art algorithms. Currently, we achieve the best score for mean squared error (MSE) on HCI 4D Light Field Benchmark.

Published
2018

49. Volumetric measurements of a self-similar adverse pressure gradient turbulent boundary layer using single-camera light-field particle image velocimetry

Author

Zhao, Zhou (author), Buchner, A.L.L. (author), Ding, Junfei (author), Shi, Shengxian (author), Atkinson, Callum (author), Soria, Julio (author), Zhao, Zhou (author), Buchner, A.L.L. (author), Ding, Junfei (author), Shi, Shengxian (author), Atkinson, Callum (author), and Soria, Julio (author)

Abstract

As a novel volumetric particle image velocimetry technique, single-camera light-field PIV (LF-PIV) is able to reconstruct three-dimensional flow fields with a single camera. The merits of LF-PIV lie in its concise hardware setup and minimum optical access requirement, its capability has been proved in many flow scenarios. In this study, LF-PIV is used to measure a self-similar adverse pressure gradient turbulent boundary layer (APG-TBL). Experiments were performed in a large water tunnel at the Laboratory for Turbulence Research in Aerospace and Combustion (LTRAC), Monash University. 20 independent batches of light-field PIV images were captured for both inner and outer flow, each consisting of 250 instantaneous image pairs. Instantaneous 3D velocity fields were reconstructed with the GPU accelerated DRT-MART and 3D cross-correlation methods and compared with two-dimensional PIV (2D-PIV) results. Initial results show that though limited by the experiment conditions and PIV algorithms developed in 2016, we still can have similar accuracy to 2D-PIV near and above the boundary layer. With the volumetric calibration method that compensates optical distortions caused by lens defect and misalignment between the micro-lens array (MLA) and image sensor, the resolution of LF-PIV is sure to have a large improvement., Fluid Mechanics

Published
2018

Catalog

Books, media, physical & digital resources
See catalog results