Generalized learning model, GLM for short, is a new kind of machine learning model which fuses symbolic learning, connective learning, fuzzy learning, evolutionary learning and statistical learning together. By introd...
详细信息
Generalized learning model, GLM for short, is a new kind of machine learning model which fuses symbolic learning, connective learning, fuzzy learning, evolutionary learning and statistical learning together. By introducing generalized learning into imagerecognition, this paper presents a new kind of imagerecognition model, GLIRM for short. The distinguished advantage of GLIRM is its adaptive learning ability. Through practical application in remotesensingimagerecognition, satisfactory results have been achieved.
Linear spectral mixture analysis (LSMA) is a widely used technique for subpxiel detection and mixed pixel classification. Due to mathematical tractability it is generally implemented without constraints. However, it h...
详细信息
ISBN:
(纸本)0819457914
Linear spectral mixture analysis (LSMA) is a widely used technique for subpxiel detection and mixed pixel classification. Due to mathematical tractability it is generally implemented without constraints. However, it has been shown that the constrained LSMA can improve the unconstrained LSMA, specifically in quantification when accurate estimates of abundance fractions are required. When the constrained LSMA is considered, two constraints are generally imposed on abundance fractions, abundance sum-to-one constraint (ASC) and abundance nonnegativity constraint (ANC), referred to as abundance-constrained LSMA (AC-LSMA). A general and common approach to solving the AC-LSMA is to estimate abundance fractions in the sense of least-squares error (LSE) subject to the imposed constraints. Since the LSE is not weighted in accordance with significance of bands, the effect caused by the LSE is assumed to be uniform over all the bands, which is generally not necessarily true. This paper extends the commonly used LSE-based AC-LSMA to weighted LSE-based AC-LSMA with the weighting matrix that is derived from various approaches such as parameter estimation, pattern classification and orthogonal subspace projection (OSP). As demonstrated by experiments, the weighted LSE-based AC-LSMA generally performs better than the commonly used LSE-based ACLSMA where the latter can be considered a special case of the former with the weighting matrix reduced to the identity matrix.
Exploring scenes using Bayesian nets is based on the idea of performing an active knowledge based search on images unlike conventional visual recognition algorithms. During the active search of images, a sample set of...
详细信息
Exploring scenes using Bayesian nets is based on the idea of performing an active knowledge based search on images unlike conventional visual recognition algorithms. During the active search of images, a sample set of training images from different classes is available right at the onset of an experiment and the nature of the class to be searched is unknown. Usually a recursive search on the images for objects, belonging to all classes is performed using a conventional object recognition system and our approach presented in the present paper can obviate this. The search by Bayesian nets can be confined only to a specific class or a set of classes, provided the relationships between constituent objects are exactly defined. We prove that if structural relationships are rightly established between the constituent objects of an image, searching scenes using Bayesian nets is quite effective and the presented results proclaim this very fact.
Fingerprint identification system is mainly consisted of fingerprint achieving, fingerprint classification and fingerprint matching. Fingerprint matching is the key to the system and effects on the precision and effic...
详细信息
Fingerprint identification system is mainly consisted of fingerprint achieving, fingerprint classification and fingerprint matching. Fingerprint matching is the key to the system and effects on the precision and efficiency of the whole system directly. Fingerprints are matched mainly based on their fingerprint texture pattern, which can be described with the orientation field of fingerprints. A fingerprint, which has the different orientation angle structure in different local area of the fingerprint and has a texture pattern correlation among the neighboring local areas of the fingerprint, can be viewed as a Markov stochastic field. A novel method of fingerprint matching, which is based on embedded hidden Markov model (HMM) that is used for modeling the fingerprint's orientation field, is described in this paper. The accurate and robust fingerprint matching can be achieved by matching embedded hidden Markov model parameters which were built after the processing of extracting features from a fingerprint, forming the samples of observation vectors and training the embedded hidden Markov model parameters.
In this paper, we propose a novel mutual information (MI) based method to register SAR and SPOT images. The traditional MI can register SAR and SPOT images well. However, its robustness is weakened by the absence of o...
详细信息
In this paper, we propose a novel mutual information (MI) based method to register SAR and SPOT images. The traditional MI can register SAR and SPOT images well. However, its robustness is weakened by the absence of orientation information. In our approach, we first extract orientation information at four directions by Gabor filters, then MI of each corresponding image pair is calculated and the average value of MI is used as an improved measure for MI. Experiments show that our method is more robust than the traditional MI method. Meanwhile our method maintains comparable accuracy to the traditional MI, which is much better than coarse manual registration.
The multiband target detection algorithms implemented in hyperspectral imaging systems represent perhaps the most successful example of image fusion. A core suite of such signal processing methods that fuse spectral c...
详细信息
The multiband target detection algorithms implemented in hyperspectral imaging systems represent perhaps the most successful example of image fusion. A core suite of such signal processing methods that fuse spectral channels has been implemented in an operational system; more systems are planned. Stricter performance requirements for future remotesensing applications will be met by evolutionary improvements on these techniques. Here we first describe the operational methods and then the related next generation nonlinear methods, whose performance is currently being evaluated. Next we show how a "dual" representation of these algorithms can serve as a springboard to a radically new direction in algorithm research. Using nonlinear mathematics borrowed from machine learning concepts, we show how hyperspectral data from a high-dimensional spectral space can be transformed onto a manifold of even higher dimension, in which robust decision surfaces can be more easily generated. Such surfaces, when projected back into spectral space, appear as enveloping blankets that circumscribe clutter distributions in a way that the standard, covariance-based methods cannot. This property may permit the design of extremely low false-alarm rate solutions to remote detection problems
This paper revises the theoretical background for upcoming dual-channel Radar satellite missions to monitor traffic from space. As it is well-known, an object moving with a velocity deviating from the assumptions inco...
详细信息
This paper revises the theoretical background for upcoming dual-channel Radar satellite missions to monitor traffic from space. As it is well-known, an object moving with a velocity deviating from the assumptions incorporated in the focusing process will generally appear both displaced and blurred in the azimuth direction. To study the impact of these (and related) distortions in focused SAR images, the analytic relations between an arbitrarily moving point scatterer and its conjugate in the SAR image have been reviewed and adapted to dual-channel satellite specifications. To be able to monitor traffic under these boundary conditions in real-life situations, a specific detection scheme is proposed. This scheme integrates complementary detection and velocity estimation algorithms with knowledge derived from external sources as, e.g., road databases.
Since PCA-based teeth-image personal identification method (K. Prajuabklang, et al., 2004) is not robust against reflection and orientation, registered persons in database are rejected around 7%. This paper proposes a...
详细信息
Since PCA-based teeth-image personal identification method (K. Prajuabklang, et al., 2004) is not robust against reflection and orientation, registered persons in database are rejected around 7%. This paper proposes a method to improve the PCA-based teeth-image personal identification method. In this method, the teeth image failed from the matching in the PCA-based system is reconsidered by feeding back the image to eliminate the reflection and the rotation problems. The enhanced teeth image is fed back to PCA process in order to rescue misclassified teeth-image. In the experiments, 25 teeth images are tested with 20-teeth database. The results revealed that of the 7% errors caused by the two problems, 5% are correctly identified because of the proposed method
暂无评论