Volume Rendering methods employing the GPU capabilities offer high performance on off-the-shelf hardware. In this article, we discuss the various bottlenecks found in the graphics hardware when performing GPU-based Vo...
详细信息
Tracing and tracking facial features in a precise manner are crucial tasks for Human computer Interaction, facial expression recognition, and image retrieval. Eyelids tracking is an important and hard subject to evalu...
详细信息
ISBN:
(纸本)0889865981
Tracing and tracking facial features in a precise manner are crucial tasks for Human computer Interaction, facial expression recognition, and image retrieval. Eyelids tracking is an important and hard subject to evaluate human emotions due to the high expressivity of the eyes and its faster movement. For Active Appearance Models (AAM) is a challenge, since related frameworks have showed a good performance using edge detectors, color information, and thresholding techniques all of them depending on the quality of the image. We built two appearance-based models (ABM) to track simultaneously eyelids tracking with 3D head pose, eyebrows and lips in monocular video sequences. this paper has two main contributions. Firstly, we show that by adopting a non-occluded facial texture model in the eyes region, 3D head pose parameters can be obtained in an accurate and stable manner. Secondly, unlike previous approaches regarding eyelids tracking, we prove that the Online Appearance Models (OAMs) can be used for eyelids tracking without color information, eyes feature extraction, or edge detectors. Experiments in real videos show the feasibility and usefulness of this approach, down-weighting time and memory consumption, and improving the accuracy.
this volume contains the latest in the series of ICAPR proceedings on the state-of-the-art of different facets of pattern recognition. these conferences have already carved out a unique position among events attended ...
详细信息
ISBN:
(数字)9789812772381
ISBN:
(纸本)9789812705532
this volume contains the latest in the series of ICAPR proceedings on the state-of-the-art of different facets of pattern recognition. these conferences have already carved out a unique position among events attended by the pattern recognition community. the contributions tackle open problems in the classic fields of image and video processing, document analysis and multimedia object retrieval as well as more advanced topics in biometrics speech and signal analysis. Many of the papers focus both on theory and application driven basic research pattern recognition.
Properties of various models of the relief used in fundamental geophysical researches of the Earth, render essential influence, on character and capabilities of developed information technologies of support of full sc...
详细信息
Properties of various models of the relief used in fundamental geophysical researches of the Earth, render essential influence, on character and capabilities of developed information technologies of support of full scale computing models of earth processes and the phenomena using space data of remote sensing. the scaled model of a relief for representation of territory of pools of the big rivers of Siberia is considered (Irtysh, Ob, Lena), allowing to make distribution and calculation of a water drain under remote data of meteorological supervision(observations), and also dynamically to update values of parameters of model with preservation of its(her) main(basic) properties in a scale number(series) of visual display of model. the main goal is - to find the decision of integral equation z = f (x, y) where minimal.
Unscented Kalman Filter (UKF) is a very popular tracking algorithm which provides a way of processing non-linear and non-Gaussian models. In this paper we propose a modified UKF to extend its applications to multi-sen...
详细信息
ISBN:
(纸本)0889865981
Unscented Kalman Filter (UKF) is a very popular tracking algorithm which provides a way of processing non-linear and non-Gaussian models. In this paper we propose a modified UKF to extend its applications to multi-sensor scenes, improving thus the global result. the combination of several independent sensors increases the precision and robustness of our tracking system, making possible to solve difficult situations, such as occlusions or serious noise. We approach this modification to multi-camera tracking applications, like video surveillance or sport performance improvement.
the subject of this paper is the development of a computervision tool, called RDV MON, for the automated visual monitoring of the rendezvous between the International Space Station (ISS) and the Automated Transfer Ve...
详细信息
ISBN:
(纸本)9290929170
the subject of this paper is the development of a computervision tool, called RDV MON, for the automated visual monitoring of the rendezvous between the International Space Station (ISS) and the Automated Transfer Vehicle (ATV). RDV MON must be capable of acquiring and processingthe video images of the rendezvous filmed by the ISS Russian Service Module camera, detecting several ATV shapes and evaluating ATV relative position and attitude, provided some information on the camera calibration parameters and on the initial conditions. Several imageprocessing techniques are used, taking into account the variability of the illumination and of the target size and shape. Algorithms have been tested using simulated rendezvous images and telemetry from the Astronauts Crew Training software (ACT-L) and pre-processed ISS/Progress rendezvous videos. the results encourage the use of computervision techniques for relative motion monitoring in the specific context of ATV/ISS rendezvous, with satisfactory performances in terms of processing time and accuracy, and show the possibility of integrating such algorithms to relative guidance and navigation.
this paper describes a method to estimate horizontal face pose from the detected facial parts using dynamic Bayesian networks (DBNs). Facial parts, which are eyes, pupils and mouth, are fully utilized to (1) construct...
详细信息
ISBN:
(纸本)0889865981
this paper describes a method to estimate horizontal face pose from the detected facial parts using dynamic Bayesian networks (DBNs). Facial parts, which are eyes, pupils and mouth, are fully utilized to (1) construct a model known as "head cylindrical model (HCM)", and (2) compute three observation values for inference using DBNs. HCM is designed originally from some anthropometric data of average pupils distance and head width. this model provides the head center, which behaves like a shaft or a joint where rotation of the head is made about this joint (similar to human head biologically). Using this model, we simplify the difficulties to compute center of human head and reduce sensitivity to noise while tracking, especially when head moves or pans. Once HCM is constructed, it will be fit into the head and kept inside it and the head will always be kept inside it along the tracking process. A scheme to maintain its position relative to face motion is also explained. the contribution from this work is that our proposed method fully utilizes detected features from tracking module and does not require any additional pre-processing stage, thus, it is computationally cheap and suitable for real-time applications. Results also reveal that our method achieves good results and performs at real-time.
Detecting manipulated images has become an important problem in many domains (including medical imaging, forensics, journalism and scientific publication) largely due to the recent success of image synthesis technique...
详细信息
Detecting manipulated images has become an important problem in many domains (including medical imaging, forensics, journalism and scientific publication) largely due to the recent success of image synthesis techniques and the accessibility of image editing software. Many previous signal-processing techniques are concerned about finding forgery through simple transformation (e.g. resizing, rotating, or scaling), yet little attention is given to examining the semantic content of an image, which is the main issue in recent image forgeries. Here, we present a complete workflow for finding the anomalies within images by combining the methods known in computergraphics and artificial intelligence. We first find perceptually meaningful regions using an image segmentation technique and classify these regions based on image statistics. We then use AI common-sense reasoning techniques to find ambiguities and anomalies within an image as well as perform reasoning across a corpus of images to identify a semantically based candidate list of potential fraudulent images. Our method introduces a novel framework for forensic reasoning, which allows detection of image tampering, even with nearly flawless mathematical techniques. (c) 2006 DFRWS. Published by Elsevier Ltd. All rights reserved.
Photomosaic is a technique which transforms an input image into a rectangular grid of thumbnail images preserving the overall appearance. the typical photomosaic algorithm searches from a large database of images one ...
详细信息
ISBN:
(纸本)8090310079
Photomosaic is a technique which transforms an input image into a rectangular grid of thumbnail images preserving the overall appearance. the typical photomosaic algorithm searches from a large database of images one picture that approximates a block of pixels in the main image. Since the quality of the output depends on the size of the database, it turns out that the bottleneck in each photomosaic algorithm is the searching process. In this paper we present a technique to speed-up this critical phase using the Antipole Tree Data Structure. this improvement allows the use of larger databases without requiring much longer processing time. Copyright UNION Agency - Science Press.
the purpose of our study is to introduce a new method for depth of field blur simulation used for compositing real and synthetic images in a realistic way. Existing methods do not simulate accurately the depth of fiel...
详细信息
ISBN:
(纸本)8090310079
the purpose of our study is to introduce a new method for depth of field blur simulation used for compositing real and synthetic images in a realistic way. Existing methods do not simulate accurately the depth of field blur produced by photographic equipment. the problem is twofold: photometric and geometric. Existing methods based on convolution do not produce a realistic photometric rendering compared to real photograph and the photographic blur contains patterns associated withthe diaphragm shape, neither handled by existing methods. Our method, based on imageprocessing, addresses these two issues by following the photographic image creation process. It can be used on synthetic or real images, essentially for special effects compositing. Copyright UNION Agency - Science Press.
暂无评论