Enhancing image quality of ghost imaging by fuzzy c-means clustering method

Abstract

We presented a novel ghost imaging scheme based on fuzzy c-means clustering (FCM) to reduce measurements and improve the visibility of the reconstruction image. Different from the GI methods, the FCM model is first employed to partition the intensity values of the reference light path and probe light path. Then, the relative speckle patterns and bucket intensity values are selected with respect to the clustering results. Finally, the object can be obtained by conventional GI methods. From the considerable simulations and experimental results, we conclude that the proposed scheme can enhance the visibility of the reconstruction image by using much fewer data from measurements compared with the existing GI methods.