Duanqing Xu

Efficient data management for incoherent ray tracing

To obtain good performance on the GPU hardware, it is necessary to design algorithms to manage data, access memory under GPU memory hierarchy, and schedule more efficient threads. In this paper, we propose an efficient data management and task management designed for GPU based ray tracing. Due to the dynamic and uncertainty in ray tracing, we design data-management layer and task-management layer combined with fuzzy spatial analysis, use the two-level ray sorting and a ray bucket structure to reorganize ray data, then a warps threads can be scheduled to access coherent geometry and nodes data, reduce memory bandwidth, and dispatch the data locally. We schedule tasks in data-driven execution according to coherent data, propose an adaptive ray compaction to eliminate inactive threads, maintain task efficiency of threads in a warp, and design two heuristics to decrease the compaction cost. On the basis of it, we also introduce a memory-optimized dynamic traversal management to reduce incoherent memory access, and avoid frequent sorting computation and compaction operations. Our experiments demonstrate all of these work combined can achieve good performance.

3D palmprint recognition using shape index representation and fragile bits

Recent years have witnessed a growing interesting in developing automatic palmprint recognition methods. Most of the previous works have concentrated on two dimensional (2D) palmprint recognition in the past decade. However, the shape information is lost in 2D plamprint images. What’s more, 2D plamprint recognition is not robust enough in practice since its data could be easily counterfeited or contaminated by noise. Consequently, three dimensional (3D) palmprint recognition is treated as an important alternative road to both enhance the performance and robustness of current available palmprint recognition systems. In this paper, we first explore geometrical information of 3D palmprint data by employing shape index formulation, from which Gabor wavelet features are then extracted. Furthermore, we first discover that by incorporating fragile bits information, the performance of coding strategy related 3D recognition method can be further improved. Experiments conducted on the public available 3D plamprint database validate that our method can obtain the highest recognition performance among the state-of-the-art methods estimated.

Ray Tracing Dynamic Scenes using fast KD-tree Base on Multi-Core Architectures

We present a simple and efficient algorithm for interactive ray tracing of dynamic Scenes using kd-tree base on multi-core architectures. Instead of alternating between phases of rendering and (potentially infrequent) building, our approach exploits the parallelism inherent in a multi-core system by continuously and asynchronously rebuilding new kd-tree (potentially over the course of multiple frames) while the remaining cores concurrently update and traverse the most recently built kd-tree. We do not make any assumptions about the possible deformation or motion of objects and dynamically update or rebuild the hierarchy depending on our simple heuristic.

Incoherent Ray tracing on GPU

Tracing secondary rays, such as reflection, refraction and shadow rays, can often be the most costly step in a modern real-time ray tracer. In this paper, we propose a new approach to ray tracing on GPU. Our approach is especially efficient for incoherent rays. Combined with the common packets ray tracing, we propose a different data-parallel approach to ray tracing on GPU, in which individual ray intersect with k different nodes/triangles in the same operation. Besides, we add some additional information in the construction of acceleration structure, and propose a new approach to travel the acceleration structure. Our acceleration structure needn’t collapse, so it could be built very efficiently, which is promising for dynamic scenes. Despite this approach is slower for primary rays, but demonstrate that it performs better than those techniques as soon as incoherent rays are considered.

Learning Local Feature Descriptors with Quadruplet Ranking Loss.

In this work, we propose a novel deep convolutional neural network (CNN) with quadruplet ranking loss to learn local feature descriptors. The proposed model receives quadruplets of two corresponding patches and two non-corresponding patches, and then outputs features measured by (L_2) norm with dimension (256d) close to that of SIFT, which thus can be easily applied to practical vision tasks by plug-in replacing SIFT-like features. Moreover, the proposed model mitigates the problem that the margin separating corresponding and non-corresponding pairs varies with samples caused by commonly used triplet loss, and improves the capacity of utilizing the limited training data. Experiments show that our model outperforms some state-of-the-art methods like TN-TG and PN-Net.

MSKD: multi-split KD-tree design on GPU

We present a novel parallel acceleration structure construction and traversal algorithm designed to efficiently exploit the massive parallel computing cores on the Graphic Processing Unit(GPU) to improve the render performance. Our associated data structure is called multi-split KD-tree or MSKD, which focuses on fast generating and efficiently traveling multiple child nodes hierarchy in parallel. At build-time, we introduce a multi-split node generation method to split along three-dimension axes into eight child nodes once, and gather quickly high-quality child nodes even at early construction phase. During traversal, we propose a progressive traversal to fast decide the visiting order for multiple child nodes. Then, we use a dynamic ray transfer to adaptively drive the traversal tasks execution on the GPU. Our experiments with this hierarchy show the construction and traversal performance improvement for ray tracing using MSKD compared to previous methods.

Complex shading efficiently for ray tracing on GPU

Complex shading often associates with long shaders and huge data access. To obtain good performance on current generation GPU hardware, it is necessary to design some algorithms to manage data, schedule more efficient threads, and memory access under the hierarchy of GPU memory. In this paper, we propose an approach to accelerate the rendering process for complex shaders by analyzing and sorting shading jobs according to their complexity and potential memory access. We show that by sorting these shading jobs in three levels of memory hierarchies and reorganizing threads block according to the complexity, all shading jobs are scheduled in order, and we can significantly improve cache utilization and GPU hardware utilization, especially for poor performance caused by large branching. All sorting work are processed on CPU with plentiful logic function, and can be processed in a very efficient manner, compared with the expensive compaction operation on GPU. Our experiments with this hierarchy demonstrate improvements against a SIMD packet tracing with compaction on GPU.