An Image Processing-Based Approach for Fuzzy Image Recognition

Developing in parallel with image processing technology, image recognition technology becomes an important sector in the artificial intelligence domain and is widely adopted in different fields. However, the existence of various fuzzy images in real life brings new challenges to image recognition. To effectively improve the feature matching rate between motion-blurred and original images, this paper first carries out a blurring simulation on the image with Gaussian noise, then performs a recovery operation on the re-blurred image by inverse filtering and Wiener filtering (WF), and later uses the ORB algorithm to perform feature point matching between the recovered image and the original image. The experimental results verify that the WF-based approach achieves satisfactory results in processing the recovery of simulated motion blur images with added noise. It is clear that the WF-based approach can contribute to the higher recognition rates motion blur image recognition.