Saliency-Based CBIR System for Exploring Lunar Surface Imagery

Abstract

Florida. His research interests include image and video retrieval, medical imaging, network and wireless communications, sensor computing, location-based services, and intelligent transportation systems. Dr. Hua has published widely, including several papers recognized as best/top papers at various international conferences. He has served as a conference chair, vice-chair, associate chair, demo chair, and program committee member for numerous conferences, and on the editorial board of the

An Image Retrieval Method for Lunar Complex Craters Integrating Visual and Depth Features

“…Based on these characteristics, similarity measurement rules and a retrieval algorithm were proposed and detailed [39]. Hua et al utilized a general saliencybased landmark detection algorithm to identify regions of interest on the lunar surface, then indexed and retrieved them using feature vectors extracted from the region-of-interest images, evaluating the performance of saliency-based landmark detection [12]. However, these methods also have apparent limitations; they perform well under specific conditions, particularly when the image content structure is simple and changes little.…”

“…Traditional CBIR methods typically rely on the visual content of images, such as texture, shape, and color features, to index and retrieve images. These methods depend on handcrafted features such as Speeded Up Robust Features (SURF) [11], Hu moments, and Gabor features [12]. Although these features are effective in certain scenarios, their application is limited in the complex lunar environment, where they struggle to capture detailed information within images effectively.…”