As one of the most successful applications of image analysis, face recognition has received significant attention, especially during past few years. Automatic human face recognition has received substantial attention ...
详细信息
As one of the most successful applications of image analysis, face recognition has received significant attention, especially during past few years. Automatic human face recognition has received substantial attention from researchers in biometrics, pattern recognition and computer vision communities. The machine learning and computergraphics communities are also increasingly involved in face recognition. The localization of human faces in digital images is a fundamental step in the process of face recognition. Although the existing automated machine recognition systems have certain level of maturity, but their accomplishments are limited due to real time challenges. For example, face recognition for the images which are acquired in high contrast with different levels of illumination is a critical problem. It is known that image variation due to lighting changes is larger than that, due to different personal identity, because lighting direction alters the relative gray scale distribution of a face image. In handling these types of practical scenarios, the system must be robust enough to deal with dynamic changes in lighting, hence it is equally important to preprocess the images prior to actual processing and experimentations. This paper proposes a novel method of illumination normalization based on histogram of an image and scaling function. It helps in construction of an optimal global lighting space from these images which improve accuracy of face recognition system. The proposed method helps in recognition of sparsely sampled images with different lighting too. Also, most valuable information of an image, i.e. gray scale value, is not discarded and person's discriminative information in face image is strengthened. Hence recognition can be carried out using preserved illumination invariant features.
TOF PET 3D reconstruction adds another layer of complexity. Thus it becomes a challenge to find the most efficient method to optimize the reconstruction without compromising image quality. In this study, mCT PET scann...
详细信息
TOF PET 3D reconstruction adds another layer of complexity. Thus it becomes a challenge to find the most efficient method to optimize the reconstruction without compromising image quality. In this study, mCT PET scanner data is used for image reconstruction and comparison. This study shows how an ultrafast forward and back projector pair TOF 3D reconstruction is implemented based on Symmetry SIMD Projector (SSP) algorithm. The Symmetry in the algorithm translates to four symmetric relationships;+90° symmetry of the projection angle, mirror-symmetry on the Radial bin axis, symmetry of the integral direction along LOR, symmetry of the oblique segment angle. The use of SIMD allows the access of four data points by one operand which reduces computation time by a factor of four. Therefore, SSP algorithm can ideally reduce the computation time of forward projection and back projection by a factor of 64. The computer hardware used is off the shelf workstation with Intel Xeon E5540 2×CPU, 12GB, 1×GPU Tesla. The reconstructed image size is 400×400×109 pixels. For the same data set, the SSP algorithm using a CPU based reconstruction was able to achieve twice the timing improvement over the standard GPU counterpart.
We developed a GPU-based real-time imaging software suite for medical ultrasound imaging to provide a fast real-time imaging platform for various probe geometries and imaging schemes. The imaging software receives raw...
详细信息
We developed a GPU-based real-time imaging software suite for medical ultrasound imaging to provide a fast real-time imaging platform for various probe geometries and imaging schemes. The imaging software receives raw RF data from a data acquisition system, and processes them on GPU to reconstruct real-time images. The most general-purpose imaging program in the suite displays three cross-sectional images for arbitrary probe geometry and various imaging schemes including conventional beamforming, synthetic beamforming, and plane-wave compounding. The other imaging programs in the software suite, derived from the general-purpose imaging program, are optimized for their own purposes, such as displaying a rotating B-mode plane and its maximum intensity projection (MIP), photoacoustic imaging, and real-time volume-rendering. Real-time imaging was successfully demonstrated using each of the imaging programs in the software suite.
The segmentation of the lung tissues in chest Computed Tomography (CT) images is an important step for developing any computer-Aided Diagnostic (CAD) system for lung cancer and other pulmonary diseases. In this paper,...
详细信息
ISBN:
(纸本)9781467364560
The segmentation of the lung tissues in chest Computed Tomography (CT) images is an important step for developing any computer-Aided Diagnostic (CAD) system for lung cancer and other pulmonary diseases. In this paper, we introduce a new framework to generate 3D realistic synthetic phantoms to validate our developed Joint Markov-Gibbs based lung segmentation approach from CT data. Our framework is based on using a 3D generalized Gauss-Markov Random Field (GGMRF) model of voxel intensities with pairwise interaction to model the 3D appearance of the lung tissues. Then, the appearance of the generated 3D phantoms is simulated based on iterative minimization of an energy function that is based on using the learned 3D-GGMRF image model. These 3D realistic phantoms can be used to evaluate the performance of any lung segmentation approach. In this paper, we used the 3D realistic phantoms to evaluate the performance of our developed lung segmentation approach based on using the Dice Similarity Coefficient (DSC) metric and the Receiver Operating Characteristics (ROC). The DSC demonstrated that our approach achieves a mean DSC value of 0.994 ± 0.0034. Moreover, the ROC analysis for our method showed the best performance (area 0.99), while intensity showed the worst performance (area 0.92).
In this paper we present a method to simulate fluids on smooth surfaces of arbitrary topology using a graphicsprocessing unit (GPU). To do this we use the parametrization of Cat mull-Clark subdivision surfaces and ob...
详细信息
In this paper, we propose a method to model the Copacabana beach sidewalk pavement, and Portuguese pavements in general. Given a black and white source image, the proposed method outputs the geometry of all individual...
详细信息
In this paper, histogram uniformization of digital images by means of the finite field cosine transform (FFCT) is examined. The approach consists in dividing the image into blocks and applying the FFCT, in a recursive...
详细信息
In video conferencing, if two people talk at the same time howling noises can appear due to failures in the echo cancellation system. In order to reduce these noise situations, a novel AEC (Acoustic Echo Cancellation)...
详细信息
image segmentation is still a challenging issue in pattern recognition. Among the various segmentation approaches are those based on graph partitioning, which present some drawbacks, one being high processing times. W...
详细信息
In this paper we propose a new and efficient texture feature extraction method: the Segmentation-based Fractal Texture Analysis, or SFTA. The extraction algorithm consists in decomposing the input image into a set of ...
详细信息
暂无评论