Improvement Of FAST Algorithm And Its Application In Computation Of Image Visual Complexity

The research and application of image-related complexity are very extensive.In addition to research areas such as automatic target recognition,image compression,image classification and retrieval,and text detection,there are also applications in medicine,cognitive psychology,and computer aesthetics.In the current studies,there is still another problem that needs to be solved.That is to give a definition of complexity that is more in line with human visual perception.Many methods that use gray scale statistics,image compression,and frequency domain transforms to calculate image complexity are not very consistent with human visual perception.In this paper,through analysis of a large number of documents,it is believed that one of the main factors that affect the visual perception of images is the feature points of the images.Finally,a weighted sum of entropy of image feature points and image colors is used to calculate the visual complexity of the image.In order to verify the feasibility and accuracy of the method,the paper compares the computational results of image visual complexity with the results of relevant literature and human visual perception scores.The results show that the method proposed in this paper is more in line with human visual experience.In addition,this paper proposes an improved FAST feature point detection algorithm.The algorithm uses the pixel similarity measurement rules of HSI color space to reduce noise and improve the accuracy of feature point detection.