Research On Iris Feature Extraction And Recognition Algorithm Based On Artificial Fish Swarm Optimization

In view of the inherent characteristics of iris texture information,such as security,reliability,unchangeable and high security,iris recognition technology has been widely concerned and studied in various fields.The technical products derived from iris recognition technology are also widely used in various industries,including security,access control,security and other systems that need to identify and verify the user's identity.Iris recognition technology has become one of the most popular biometrics authentication technologies.Since there is no unified standard for iris image collection environment,iris images collected will be different from each other.Therefore,iris images taken in an ideal constrained state are selected as research objects in this paper to improve the accuracy of traditional iris recognition.On the premise of traditional coding recognition mode,artificial fish swarm algorithm is used to perform adaptive optimization of Gabor filter parameters to extract iris features,and SVM support vector machine based on random forest is used to improve the iris recognition algorithm.In this way,the feature extraction and recognition of iris recognition system can be optimized to improve optimization performance of iris recognition algorithm.The main contents of this paper include the following:(1)for image collection: the environment of iris image collection has a great influence on the quality of iris image collection.Therefore,when iris image collection is carried out,the collection state of the collected person is specified,so that the qualified iris image can be collected under the specified state.Namely,the iris image taken under the constraint ideal state.(2)iris image pre-processing: the pre-processing mainly includes the quality evaluation of the collected iris image,the positioning of the inner and outer circle boundary of iris,the normalization of iris image,the normalization of image enhancement and the interception of ROI region.Through this series of pretreatment,the overall recognition efficiency of iris recognition system is improved.(3)iris feature extraction: the adaptive Gabor filtering iris feature extraction algorithm based on artificial fish swarm algorithm proposed in this paper shows that different Gabor parameters have a great impact on iris recognition results.Therefore,the artificial fish swarm algorithm is used to optimize the parameters of Gabor filter,so that the filter can adjust the parameters adaptively to optimize the feature extraction of different iris library images.The texture extracted by the filter is converted into binary code to facilitate subsequent recognition.(4)for iris recognition: the texture extracted by the filter usually contains some noise and redundancy,and the uniqueness of each feature coding cannot be guaranteed.Therefore,in this paper,random forest is used for feature selection to accurately extract effective features.Then the acquired feature codes were input into SVM support vector machine for training,and the trained model was used for iris feature recognition,so as to improve the recognition performance of the overall iris recognition system.In this paper,iris images from CASIA iris library of institute of automation,Chinese academy of sciences and JLU iris library of biometrics and information security technology laboratory,Jilin university were selected as experimental data.By calculating the equierror rate,the correct recognition rate and the ROC curve as the experimental performance criteria of the proposed algorithm,the optimization performance of the proposed algorithm on feature extraction and feature recognition,and the optimization effect of the proposed algorithm on the overall iris recognition system are analyzed.In summary,this paper USES artificial fish swarm to optimize Gabor filter parameters,obtains an adaptive Gabor filter feature extraction method,and converts the extracted features into the corresponding binary code.Through random forest screening feature coding and combining SVM for iris recognition,the optimization degree of iris recognition system algorithm is finally demonstrated by experiments.