Application Of Information Entropy Ant Algorithm On Feature Selection And Image Recognition

The method of swarm intelligence is a new technology to solve most global optimization problems effectively. As a realization of colony intelligence, the ant colony algorithm has been paid extensive attention by academe. Now the application fields of swarm intelligence have extended to multi-goals optimization, data classification and clustering, biology system modeling, imitation and system identification, etc. and the theory and method of swarm intelligence offer a new path to solve these application problems. So it has important academic significance and practical value to develop the research of ant swarm intelligence theories and applications.The methods of ant colony algorithm are researched deeply to apply to the two aspects including feature selection and image recognition. The efficiency of the algorithms in applications is proved by simulation experiments. The main contents of the paper are as follow.Features selection plays an important role in the image recognition field and affects the correct rate and speed rate of image recognition directly. Selecting correct and effective features has become the main procedure in image recognition. In order to select features which can detach all kinds of samples as far as possible, a feature selection algorithm is proposed by the improved ant colony algorithm based on the entropy. Eigenvectors of multiple images are inputted to the ant colony recognition algorithm at first, arbitrary characteristics of each image are as a group of the cluster center. Then move ants as the rule, change the information of regional characteristics until the termination conditions is satisfied. In the end, attain to the maximum value in the objective function, feature of which is as the best choice of features.An image recognition algorithm is proposed by the improving Information Entropy ant algorithm based on the entropy, used effectively the ant colony algorithm optimization capabilities, and realized image recognition. The method is to set up the database for all samples at first. And then, match features between extractions the features and image recognition template in the end, calculate value of the information in each route, the shorter route the higher value of the information, choose the maximum value of information addition as the best recognition result. Experimental results indicate the algorithm has higher recognition rate to standard digital identification.