Study On Application Of Image Recognition Based On Ant Colony Algorithm

As a new intelligent bionic model, ant colony algorithm is a heuristic search algorithm to solve combinatorial optimization problems. Because its many unique intelligent characters such as intelligent searching, global optimization, parallel calculation, positive feedback, steadiness, easy combined with other algorithms and so forth, ant colony algorithm has received extensive attention from academe since raised and widely used in various fields. Not only shows unique advantages to solve complex optimization problems in working out combinatorial optimization problems, such as the traveling salesman problem, quadratic assignment, workshop task scheduling and other, but also has been explored and applied research in other areas including multi-objective optimization, data classification, system modeling, simulation system identification, etc, meanwhile offers a new path to solve practical application problem.In the field of image processing, ant colony algorithm has achieved preliminary research results in term of image segmentation, image edge detection, image classification, proved the feasibility and effectiveness of ant colony algorithm in image processing applications, but the study results shows that the efficiency of ant colony algorithm for solving practical problems of image processing is not high and the computing time is longer, which duing to the complexity of the ant colony algorithm model itself and image processing, at the same time, there has not yet been found to identify a widely applied model in the image recognition, so there is large research space and requires people further study.The work of this task aims to apply and research ant colony algorithm in image recognition includes follow aspects: First, researching and summing up the basic principles and mathematical models of ant colony algorithm, analyzing deeply the advantages and disadvantages of the algorithm, improving exploration the algorithm for its shortcomings, summarizing the progressing application of ant colony algorithm in terms of image recognition, which provides a theoretical basis for subsequent image recognition applications. Second, in allusion to the image feature extraction as the key link of image recognition processing accomplish the fundamental theory research, propose pulse coupled neural network algorithm, determine a more comprehensive method of extracting characteristic function, make up the shortcomings of the ant colony algorithm. Finally, according to the ant colony with its own characteristics, combined with pulse coupled neural network, proposed a new pulse coupling-ant colony clustering algorithm and applied to the field of medical image recognition, the experimental results show that this algorithm is effective and applicable.