Hongrui Jiang's research while affiliated with University of Wisconsin–Madison and other places

... Recent works in NeRF and implicit surface reconstruction utilize Marching Cubes, [32], [33], which employ a volumetric representation to apply a discrete cuboid-shape grid and evaluate the mesh passing through each cube using zeros and triangles [32]. In cases of PMVS or explicit surface reconstruction, Delaunay triangulation is deployed [34], [35], to merge 3D points into triangles based on certain rules. On that note in [36], a global 3D mesh is constructed utilizing volumetric representation, with the additional provision of optional semantic annotations through 2D pixel-wise segmentation and 3D Bayesian updates. ...

... This type of sensor can provide comprehensive and accurate mechanical information that equals to or even exceeds human tactility (1,2). Meanwhile, recognition of multipoint pressure distribution by sensor array allows us to obtain a complete pressure interaction profile (3)(4)(5), and complemented by artificial intelligence and data analytics, researchers already show the feasibility of more intelligent sensing functions such as grasping recognition (6)(7)(8) and slip detection (2,9). In addition, the thin and lightweight sensor form is also essential, as this contributes to the feasibility and flexibility of system-level integration (10)(11)(12). ...

... The paper optimizes for mobile GPU and runs on Jetson nano and achieves 81 fps for HD videos. Similarly, [13] performs stabilization in 30 ms on GTX2080 GPU for 720p videos by use of SURF features on GPU. [33] incorporates feature based tracking and piece-wise scene planarity modeling to warp consecutive frames in a content preserving manner. ...

... Among the several optical transduction mechanisms, the use of the localized surface plasmon resonance (LSPR) phenomenon, related to the interaction of light with the free electrons of noble metal nanoparticles [10][11][12][13][14][15][16][17][18], has already revealed its capabilities in (bio)chemical sensing [19][20][21][22] and gas sensing [23][24][25][26]. Moreover, regarding the detection of physical and mechanical stimuli, such as temperature and force, some pioneering works have been published, most of them related to theoretical approaches in ultra-thin gold films [27][28][29][30][31][32][33][34][35][36][37][38] and arrays of plasmonic nanoparticles [39][40][41][42][43][44][45][46][47][48]. These and other works led gold films [27][28][29][30][31][32][33][34][35][36][37][38] and arrays of plasmonic nanoparticles [39][40][41][42][43][44][45][46][47][48]. ...

... Because MVV data are obtained from cameras that are located on the horizontal and vertical planes, inter-view correlation could be further exploited for artifact reduction [18], [19]. In [20], a multiview denoising network (MVCNN) was presented to manage 3D focus image stacks and fuse them to compute an enhanced target image. The MVCNN was trained with residual learning and batch normalization on top of a denoising CNN [21]. ...

... The development and optimization of artificial retina provide a new possibility for the treatment of retinal degeneration and irreversible damage caused by various causes [136][137][138]. Due to the biological characteristics of nanomaterials, they are increasingly used in various structures and links involved in retinal light signal conversion, in order to develop more sensitive, durable, and stable artificial retina [139][140][141]. NPs can be used to form artificial photoreceptors in the artificial retina. ...

... Moreover, the wide view and the capability to explore multiple angles of view resulted in achieving the necessary exposure for the CVS significantly faster than what was achieved in the standard scope group. 20 (100.0%) > 2 days 2 (9.5%) 0 (0.0%) Same admission complication-n (%) 1 2 (9.5%) 2 (9.5%) ...

... In the case of compound eyes, there are models with a hemispherical structure formed by polymer microlenses in honeycomb arrangements, with substrates or holes, and with spherical, cylindrical (SELFOC type), and free-form geometries, regularly placed on curved surfaces [23][24][25][26][27][28][29][30]. On the other hand, some models contemplate using photodetectors or image sensors [6,11,31] and hybrid models [21]. ...

... [23] In 2019, Van Grinsven et al. presented a liquid lens actuated by the Lorentz force, whose corresponding focal length can vary with the working current. [24] In 2020, Jamali et al. developed a liquid crystal lens for presbyopia correction, and its diopter successfully changes from 1.5 D to 0 D. [25] These electrically or mag-netically actuated variable focus liquid lens have advantages of no mechanical movement, simple structure, low cost, and high compatibility. [21][22][23][24][25] However, these liquid-filled lenses are prone to leakage, evaporation and fluidity, resulting in unstable lens performance. ...

... The quality of the measurement can be further improved by changing the coil structure, integrating a ferrite magnet, or using inductive sensors [27][28][29]. Nevertheless, it is also possible to implant capacitive biosensors in other regions of the eye, like the sclera, vitreous humor, retina, or iris [30][31][32][33][34]. Though, biomechanical sensors typically lack the accuracy of fit to an individual patient, they are already utilized to measure IOP. ...