Though hyperspectral imaging (HSI, or hypercube) has been applied in a wide range of applications, it suffers from massive volume of data for efficient data storage and transmission. Herein improved 3D discrete cosine...
详细信息
Determining the direction of arrival (DOA) of a sound source is important in spatial audio signalprocessing, as it can lead to substantial improvement in noise reduction performance. Techniques such as generalized cr...
详细信息
Classification of micro calcification clusters is very essential for early detection of breast cancer from mammograms. In this paper, an improved support vector machine (SVM) scheme is proposed, where optimized decisi...
详细信息
Though hyperspectral imaging (HSI, or hypercube) has been applied in a wide range of applications, it suffers from massive volume of data for efficient data storage and transmission. Herein improved 3D discrete cosine...
详细信息
Though hyperspectral imaging (HSI, or hypercube) has been applied in a wide range of applications, it suffers from massive volume of data for efficient data storage and transmission. Herein improved 3D discrete cosine transform (DCT) is proposed with good results yielded. Experiments on several datasets have validated the efficacy of our approach.
Three novel micro-Doppler feature extraction algorithms are presented and applied to a dataset containing real X-band radar data of moving ground targets. In each case data dimensional reduction was carried out using ...
详细信息
ISBN:
(纸本)9781629938431
Three novel micro-Doppler feature extraction algorithms are presented and applied to a dataset containing real X-band radar data of moving ground targets. In each case data dimensional reduction was carried out using principal component analysis (PCA) and incorporated into the feature extraction process. Extracted features are classified using a support vector machine (SVM) classifier. It was found that all three algorithms were able to produce classification accuracies in excess of 90%. The performance of the different algorithms are shown to depend on the method used and the degree of dimensionality reduction imposed at the PCA stage.
Classification of micro calcification clusters is very essential for early detection of breast cancer from mammograms. In this paper, an improved support vector machine (SVM) scheme is proposed, where optimized decisi...
详细信息
Classification of micro calcification clusters is very essential for early detection of breast cancer from mammograms. In this paper, an improved support vector machine (SVM) scheme is proposed, where optimized decision making is introduced for effective and more accurate data classification. Experimental results on the well-known DDSM database have shown that the proposed method can significantly increase the performance in terms of F1 and Az measurements for the successful classification of clustered micro calcifications.
Three novel micro-Doppler feature extraction algorithms are presented and applied to a dataset containing real X-band radar data of moving ground targets. In each case data dimensional reduction was carried out using ...
详细信息
ISBN:
(纸本)9781849197748
Three novel micro-Doppler feature extraction algorithms are presented and applied to a dataset containing real X-band radar data of moving ground targets. In each case data dimensional reduction was carried out using principal component analysis (PCA) and incorporated into the feature extraction process. Extracted features are classified using a support vector machine (SVM) classifier. It was found that all three algorithms were able to produce classification accuracies in excess of 90%. The performance of the different algorithms are shown to depend on the method used and the degree of dimensionality reduction imposed at the PCA stage.
This paper presents a feature extraction method known as Pulse Active Harmonic (PAH) implemented on electrocardiograph (ECG) signals for biometric authentication. The technique is based on a pulse width modulation con...
详细信息
This paper presents a feature extraction method known as Pulse Active Harmonic (PAH) implemented on electrocardiograph (ECG) signals for biometric authentication. The technique is based on a pulse width modulation concept. A total of 200 Eggs from 100 subjects, taken from the Physikalisch-Technische Bundesanstalt (PTB) database are used in the simulations. Biometric performance profile such as the area under ROC (AUR) and equal error rate (EER) are then used to evaluate the results. This work shows that: a) the PAH method outperforms conventional temporal feature extraction techniques, b) its performance is comparable to other, previously introduced, Pulse Active based methods. Hence, owing to their close similarities Pulse Active approaches can readily be combined to form robust and inherently secure biometric authentication systems.
Distance measurement is a quantitative tool to measure similarity or dissimilarity between two objects. A correct selection of distance measurement will enhance the performance of a biometric authentication system. In...
详细信息
Distance measurement is a quantitative tool to measure similarity or dissimilarity between two objects. A correct selection of distance measurement will enhance the performance of a biometric authentication system. In this paper, various types of distance measurement methods such as Euclidean distance, City block distance, Chebyshevdistance, Minkowskidistance, Sorensen distance, Cosine distance and Mahalanobis distance are evaluated to determine the best similarity measure to be used with the novel pulse active ratio (PAR) feature extraction method. The results are obtained based on comparing 486 electrocardiography (ECG)signals which provide a total of 42,149 ECG comparisons. The comparisons show that the similarity measurement based on Sorensen distance gives the best matching algorithm to increase the performance of the PAR ECG biometric approach.
暂无评论