In this paper, a discriminative structured dictionary learning algorithm is presented. To enhance the dictionary's discriminative power, the reconstruction error, classification error and inhomogeneous representat...In this paper, a discriminative structured dictionary learning algorithm is presented. To enhance the dictionary's discriminative power, the reconstruction error, classification error and inhomogeneous representation error are integrated into the objective function. The proposed approach learns a single structured dictionary and a linear classifier jointly. The learned dictionary encourages the samples from the same class to have similar sparse codes, and the samples from different classes to have dissimilar sparse codes. The solution to the objective function is achieved by employing a feature-sign search algorithm and Lagrange dual method. Experimental results on three public databases demonstrate that the proposed approach outperforms several recently proposed dictionary learning techniques for classification.展开更多
Monomorphic birds cannot be sexed visually and discriminant functions on the basis of external morphological variations are frequently used. Our objective was to evaluate the reliability of sex classification function...Monomorphic birds cannot be sexed visually and discriminant functions on the basis of external morphological variations are frequently used. Our objective was to evaluate the reliability of sex classification functions created from structural measurements of Chilean flamingos Phoenicopterus chilensis museum skins for the gender assignment of live birds. Five measurements were used to develop four discriminant functions： culmen, bill height and width, tarsus length and middle toe claw. The functions were tested on a sample of live flamingos from a zoo. The best classification for museum flamingos was given by a function using tarsus length, bill width and middle toe claw （97%）. However, this function did not give the best classification for the zoo-based flamingos （81%） which had the best sex assignment by a function including measurements of tarsus, culmen and bill height and width （85%）. This shows that a function giving good results in the sample from which it originated may not be as good when applied to another group of animals. Our study emphasizes the need for assessing the accuracy of a function by testing it with other methods to ensure its suitability when being applied .展开更多
基金Supported by the National Natural Science Foundation of China(No.61379014)
文摘In this paper, a discriminative structured dictionary learning algorithm is presented. To enhance the dictionary's discriminative power, the reconstruction error, classification error and inhomogeneous representation error are integrated into the objective function. The proposed approach learns a single structured dictionary and a linear classifier jointly. The learned dictionary encourages the samples from the same class to have similar sparse codes, and the samples from different classes to have dissimilar sparse codes. The solution to the objective function is achieved by employing a feature-sign search algorithm and Lagrange dual method. Experimental results on three public databases demonstrate that the proposed approach outperforms several recently proposed dictionary learning techniques for classification.
文摘Monomorphic birds cannot be sexed visually and discriminant functions on the basis of external morphological variations are frequently used. Our objective was to evaluate the reliability of sex classification functions created from structural measurements of Chilean flamingos Phoenicopterus chilensis museum skins for the gender assignment of live birds. Five measurements were used to develop four discriminant functions： culmen, bill height and width, tarsus length and middle toe claw. The functions were tested on a sample of live flamingos from a zoo. The best classification for museum flamingos was given by a function using tarsus length, bill width and middle toe claw （97%）. However, this function did not give the best classification for the zoo-based flamingos （81%） which had the best sex assignment by a function including measurements of tarsus, culmen and bill height and width （85%）. This shows that a function giving good results in the sample from which it originated may not be as good when applied to another group of animals. Our study emphasizes the need for assessing the accuracy of a function by testing it with other methods to ensure its suitability when being applied .
