Both face detection and face recognition have started to be used widely these days in various applications such as biometric, surveillance, security, advertisement, entertainment, and so on. The ever increasing input ...
详细信息
ISBN:
(纸本)9781479963089
Both face detection and face recognition have started to be used widely these days in various applications such as biometric, surveillance, security, advertisement, entertainment, and so on. The ever increasing input image size in face detection and the large input DB in face recognition keep requiring more computational power to achieve real-time processing. Recently, embedded GPUs have started to support OpenCL and many applications can be accelerated successfully as the server GPUs have. In this paper, we propose several optimization techniques for the Local Binary Pattern (LBP) based integrated face detection and recognition algorithms, and successfully accelerated them achieving 22 fps using OpenCL on ARM Mali GPU, and 38 fps using CUDA on Tegra K1 GPU for HD inputs. This corresponds to 2.9 times and 3.7 times speedups respectively. To the best of our knowledge, it is the first paper that presents the acceleration of the face detection on embedded GPGPUs, and also that presents the performance of Tegra K1 GPU.
This work introduces a residual operator called ultimate attribute leveling. We also present an efficient algorithm for ultimate attribute leveling computation by using a structure called tree of shapes. Our algorithm...
详细信息
In this paper, we present a tool, called 3DintEx, that allows interactive exploration of the structural and behavioral aspects of a system model in 3D environments. The tool accepts a CAD model of the system and other...
详细信息
In this paper, we present a tool, called 3DintEx, that allows interactive exploration of the structural and behavioral aspects of a system model in 3D environments. The tool accepts a CAD model of the system and other data sets that describe the system safety behavioral aspects. 3DintEx was built using the Vrui framework [6], which makes it portable and scalable on a range of VR environments (e.g., desktop 3D displays, PowerWall, or CAVE). 3DintEx extends the Vrui framework in order to provide options that ease up the interactive exploration of the CAD model. Moreover, it was designed to link the available input devices to 3D interaction facilities for extracting more information from the system 3D model.
In this paper we introduce a novel approach for stereoscopic rendering of virtual environments with a wide Field-of-View (FoV) up to 360°. Handling such a wide FoV implies the use of non-planar projections and ge...
详细信息
In this paper we introduce a novel approach for stereoscopic rendering of virtual environments with a wide Field-of-View (FoV) up to 360°. Handling such a wide FoV implies the use of non-planar projections and generates specific problems such as for rasterization and clipping of primitives. We propose a novel pre-clip stage specifically adapted to geometric approaches for which problems occur with polygons spanning across the projection discontinuities. Our approach integrates seamlessly with immersive virtual reality systems as it is compatible with stereoscopy, head-tracking, and multi-surface projections. The benchmarking of our approach with different hardware setups could show that it is well compliant with real-time constraint, and capable of displaying a wide range of FoVs. Thus, our geometric approach could be used in various VR applications in which the user needs to extend the FoV and apprehend more visual information.
In this paper we introduce the Reduction Sweep algorithm, a novel graph-based image segmentation algorithm that is designed for easy parallelization. It is based on a clustering approach focusing on local image charac...
详细信息
The Finite Difference Time Domain (FDTD) method is popular for solving Maxwell's equations in time domain. The FDTD domains are usually open regions. For practical applications, since no computer can store infinit...
详细信息
The Finite Difference Time Domain (FDTD) method is popular for solving Maxwell's equations in time domain. The FDTD domains are usually open regions. For practical applications, since no computer can store infinite data, the FDTD method had to introduce Absorbing Boundary Conditions (ABCs). The ABCs should be able to absorb the outgoing waves and simulate the extension to infinity. The Complex Frequency-Shifted Perfectly Matched Layer (CFS-PML) ABCs (M. Kuzuoglu and R. Mittra. Frequency Dependence of the Constitutive Parameters of Causal Perfectly Matched Absorbers. IEEE Micr. Guid. Wave Let., 6(12):447449, Dec. 1996) has been the most widely used and effective ABC since it perfectly absorbs the outgoing waves. However, the implementation of the CFS-PML is complicated and requires more computational resources. The Huygens Absorbing Boundary Condition (HABC) (J.-P. Bérenger. On the Huygens Absorbing Boundary Conditions for Electromagnetics. J. Comp. Phys., 226:354378, 2007) was recently introduced by Jean-Pierre Bérenger. The HABC is simpler to implement than the CFS-PML. Also, the HABC is proofed to be comparable in accuracy with the CFS-PML.
Capturing and creating models of human subjects is an ongoing effort in computergraphics. In this paper, we present a framework to capture, model, simplify, and stylize an human subject. We use a consumer 3D depth ca...
详细信息
Modeling organic objects and natural phenomena is a very time consuming task, and usually requires a certain level of expertise. This happens due to the lack of suitable geometric primitives available in most modeling...
详细信息
Among various image retrieval approaches, the use of sketches lets one express a precise visual query with simple and widespread means. The challenge consists in finding a content representation that allows you to eff...
详细信息
In this paper we propose a visual estimation algorithm to aid the guidance system of an unmanned aerial vehicle (UAV). Specifically, we are considering a glider UAV trying to reach a target zone. Given that the UAV is...
详细信息
暂无评论