Image Processing Based On Bacterial Foraging Optimization Algorithm

With the growing developments of computer technology and updating of hardware and software,more and more people process images in various ways using computers.Image processing technology,such as machine vision and industry detection,has been developed rapidly and the range of its application has been continuously expanded.It is a remarkable subject with broad prospects involving the technologies of image segment,image enhancement,image classification and object recognition,etc.The following works of image processing can do well if Image segmentation and image classification have been handledproperly,otherwise it is difficult to do the following works,such as object recognition.However,basic algorithms of image segmentation and image classification only solve a part of some problems.And image segmentation algorithms and image classification algorithms based on bionic algorithms can further solve some problems.However,tradition bionic algorithms have weakness such as local optimum,poor efficiency and low search precision.So an improved bacterial foraging optimization algorithm has been proposed in this paper for image segmentation and image classification through the analysis,improvement and applications of bacterial foraging optimization,main research as follow.(1)An improved bacterial foraging optimization(IBFO)is proposed and applied to segment images.The elimination-dispersal operation and chemotaxis operation of BFO algorithm are adjusted dynamically to solve the problems that the accuracy and the efficiency are not high while processing large numbers of images using it,and bacterial individual can change the way according to its stage.IBFO is used for segmenting images.The experimental results prove that the effectiveness of improved bacterial foraging optimization proposed in this paper is better than that of other swarm intelligence algorithms by using indicators such as functional convergence,regional gray level,non-uniformity and time consumption,etc.(2)The second improved bacterial foraging optimization(SIBFO)is proposed and applied to classify object images.This improved bacterial foraging optimization is used to classify these objects in images such as pedestrians,cars and pets extracted from videos.Shown from some indicators such as recall rate,precision rate and weighed indicator combined,we can know that the effectiveness of the object classification method based on the improved bacterial foraging optimization proposed in this paper is better than those of other traditional swarm intelligence algorithms.(3)A parameters optimization method of SVM based on SIBFO is proposed and the optimal SVM is used for face recognition.The optimal classifier is obtained to recognize facial images.Shown from the indicators such as global searching,predicted precision and error analysis,classification accuracy of face recognition on ORL dataset and AR dataset,we can know that the effectiveness of parameters optimization method based on the improved bacterial foraging optimization proposed in this paper is better than those of other traditional swarm intelligence algorithms.