The paper presents a novel probabilistic framework for 3D surface reconstruction from multiple stereo images. The method works on a discrete voxelized representation of the scene. An iterative scheme is used to estima...
详细信息
ISBN:
(纸本)0769512720
The paper presents a novel probabilistic framework for 3D surface reconstruction from multiple stereo images. The method works on a discrete voxelized representation of the scene. An iterative scheme is used to estimate the probability that a scene point lies on the true 3D surface. The novelty of our approach lies in the ability to model and recover surfaces which may be occluded in some views. This is done by explicitly estimating the probabilities that a 3D scene point is visible in a particular view from the set of given images. This relies on the fact that for a point on a lambertian surface, if the pixel intensities of its projection along two views differ, then the point is necessarily occluded in one of the views. We present results of surface reconstruction from both real and synthetic image sets.
This paper presents a result of basic research with an interactive sound presentation system to aim at making human perceive color imagepatterns as a goal. The system has the following characteristics. (1) Operators ...
详细信息
This paper presents a result of basic research with an interactive sound presentation system to aim at making human perceive color imagepatterns as a goal. The system has the following characteristics. (1) Operators can interactively input positions of interest to be heard through pen-tablet interfaces. (2) The system presents local colored patterns by the form of concatenated syllables, i.e., a kind of "color pattern verbal words". (3) An effect of learning is studied on correct perceptual rate. (4) The positional assignments of syllables to the color pattern verbal words are examined, considering human perceptual characteristics.
The article deals with the recovery of illusory linear clues from perspectively skewed documents, with the purpose of using them for rectification. The computational approach proposed implements the perceptual organiz...
详细信息
ISBN:
(纸本)0769512720
The article deals with the recovery of illusory linear clues from perspectively skewed documents, with the purpose of using them for rectification. The computational approach proposed implements the perceptual organization principles implicitly used in textual layouts. The numerous examples provided show that the method is robust and viewpoint and scale invariant.
Video-based eye gaze detection systems are useful for eye-slaved support systems for the severely disabled. The pupil center in the video image is a focal point to determine the eye gaze. Recently, to improve the disa...
详细信息
Video-based eye gaze detection systems are useful for eye-slaved support systems for the severely disabled. The pupil center in the video image is a focal point to determine the eye gaze. Recently, to improve the disadvantages of traditional pupil detection methods, a pupil detection technique using two light sources (LEDs) and the image difference method was proposed. In addition, for users or subjects wearing corrective eyeglasses a method for eliminating the images of the light sources reflected in the glass lens was proposed. However, image-processing.hardware for implementing these methods is rather expensive. In the present paper, the hardware construction is replaced by a construction consisting of a combination of a conventional image grabber and a personal computer. An algorithm for windowing around the pupil image with an automatic thresholding method for pupil detection is proposed. The results show that the algorithm works well when the user or the subject is wearing eyeglasses and under normal ambient lighting conditions. The calculation time is quick enough for real time processing. These algorithms would contribute to consistent and reliable pupil detection.
Face recognition has become an important topic within the field of patternrecognition and computer vision. In this field a number of different approaches to feature extraction, modeling and classification techniques ...
详细信息
Face recognition has become an important topic within the field of patternrecognition and computer vision. In this field a number of different approaches to feature extraction, modeling and classification techniques have been tested. However, many questions concerning the optimal modeling techniques for high performance face recognition are still open. The face recognition system developed by our research group uses a discrete cosine transform (DCT) combined with the use of pseudo-2D hidden Markov models (P2DHMM). In the past our system used continuous probability density functions and was tested on a smaller database. This paper addresses the question of the presence of a major difference in recognition performance with discrete production probabilities compared to continuous ones. Therefore the system is tested using a larger subset of the FERET database. We show that we are able to achieve higher recognition scores and an improvement concerning the computation speed by using discrete modeling techniques.
A new method of feature extraction with a rotation invariant property is presented. One of the main contributions of this study is that a rotation invariant signature of 2D contours is selected based on fractal theory...
详细信息
A new method of feature extraction with a rotation invariant property is presented. One of the main contributions of this study is that a rotation invariant signature of 2D contours is selected based on fractal theory. The rotation invariant signature is a measure of the fractal dimensions, which is rotation invariant based on a series of central projection transform (CPT) groups. As the CPT is applied to a 2D object, a unique contour is obtained. In the unfolding process, this contour is further spread into a central projection unfolded curve, which can be viewed as a periodic function due to the different orientations of the pattern. We consider the unfolded curves to be non-empty and bounded sets in IR/sup n/, and the central projection unfolded curves with respect to the box computing dimension are rotation invariant. Some experiments with positive results have been conducted. This approach is applicable to a wide range of areas such as image analysis, patternrecognition etc.
To utilize CT or MRI images for computer aided diagnosis applications, robust features that represent 3D image data need to be constructed and subsequently used by a classification method. We present a computer aided ...
详细信息
ISBN:
(纸本)0769512720
To utilize CT or MRI images for computer aided diagnosis applications, robust features that represent 3D image data need to be constructed and subsequently used by a classification method. We present a computer aided diagnosis system for early diagnosis of colon cancer. The system extracts features via a new 3D patternprocessing.method and processes them using a support vector machine classifier. Our 3D patternprocessing.method, called Random Orthogonal Shape Section (ROSS) mimics the radiologist's way of viewing these images and combines information from many random triples of mutually orthogonal sections going through the volume. Another contribution of the paper is a new feedback framework between the classification algorithm and the definition of the features. This framework, called Distinctive Component Analysis combines support vector samples with linear discriminant analysis to map the features of clustered support vectors to a lower dimensional space where the two classes of objects of interest are optimally separated to obtain better features. We show that the combination of these better features with support vector machine classification provides a good recognition rate.
Fingerprint recognition and verification are often based on local fingerprint features, usually ridge endings or terminations, also called minutiae. By exploiting the structural uniqueness of the image region around a...
详细信息
Fingerprint recognition and verification are often based on local fingerprint features, usually ridge endings or terminations, also called minutiae. By exploiting the structural uniqueness of the image region around a minutia, the fingerprint recognition performance can be significantly enhanced. However, for most fingerprint images the number of minutia image regions (MIRs) becomes dramatically large, which imposes - especially for embedded systems - an enormous memory requirement. Therefore, we are investigating different algorithms for compression of minutia regions. The requirement for these algorithms is to achieve a high compression rate (about 20) with minimum loss in the matching performance of minutia image region matching. We investigate the matching performance for compression algorithms based on the principal component and the wavelet transformation. The matching results are presented in form of normalized ROC curves and interpreted in terms of compression rates and the MIR dimension.
This paper suggests an automatic diagnostic system for esophageal motility records using neural networks. Signal processing.techniques, feature extraction, and patternrecognition criteria were combined to develop com...
详细信息
This paper suggests an automatic diagnostic system for esophageal motility records using neural networks. Signal processing.techniques, feature extraction, and patternrecognition criteria were combined to develop computer programs to be used in identifying, characterizing and classifying of esophageal motility recordings. The architecture of such an automated system includes four cooperating modules: a digital filter to remove the interfered noise, separation of peristaltic waveforms from the tubular region of the esophagus, feature extraction module to detect the main quantitative parameters of each esophageal part, and a multilayer feed-forward neural network trained using the conjugate gradient algorithm was used to classify the peristalsis into different categories. The percentage of correct classification reaches 100%.
Appearance-based object recognition systems rely on training from imagery, which allows the recognition of objects without requiting a 3D geometric model. It has been little explored whether such systems can be traine...
详细信息
ISBN:
(纸本)0769512720
Appearance-based object recognition systems rely on training from imagery, which allows the recognition of objects without requiting a 3D geometric model. It has been little explored whether such systems can be trained from imagery that is unlabeled, and whether they can be trained from imagery that is not trivially segmentable. In this paper we present a method for minimally supervised training of a previously developed recognition system from unlabeled and unsegmented imagery. We show that the system can successfully extend an object representation extracted from one black background image to contain object features extracted from unlabeled cluttered images and can use the extended representation to improve recognition performance on a test set.
暂无评论