An Unsupervised Multicode Hashing Method for Accurate and Scalable Remote Sensing Image Retrieval

Abstract

Hashing methods have recently attracted great attention for approximate nearest neighbor search in massive remote sensing (RS) image archives due to their computational and storage effectiveness. The existing hashing methods in RS represent each image with a single-hash code that is usually obtained by applying hash functions to global image representations. Such an approach may not optimally represent the complex information content of RS images. To overcome this problem, in this letter, we present a simple yet effective unsupervised method that represents each image with primitive-cluster sensitive multi-hash codes (each of which corresponds to a primitive present in the image). To this end, the proposed method consists of two main steps: 1) characterization of images by descriptors of primitive-sensitive clusters and 2) definition of multi-hash codes from the descriptors of the primitive-sensitive clusters. After obtaining multi-hash codes for each image, retrieval of images is achieved based on a multi-hash-code-matching scheme. Any hashing method that provides single-hash code can be embedded within the proposed method to provide primitive-sensitive multihash codes. Compared with state-of-the-art single-code hashing methods in RS, the proposed method achieves higher retrieval accuracy under the same retrieval time, and thus it is more efficient for operational applications.

Remote Sensing Image Retrieval in the Past Decade: Achievements, Challenges, and Future Directions

“…Chen et al have proposed an unsupervised multispectral RSIR method, making use of the unsupervised representation learning ability of GAN (Generative Adversarial Network) [92]. Reato et al have presented a simple yet effective unsupervised RSIR method that represented each image with primitive-cluster sensitive multi-hash codes [93]. Lukac et al have improved the well-known KLSH (kernelized locality sensitive hashing) method using Graphical Processing Units (GPU) to make it feasible for parallelization, and thus performing fast parallel image retrieval [94].…”

A Deep Cross-Modality Hashing Network for SAR and Optical Remote Sensing Images Retrieval

“…Moreover, both deep semantic features and weighted distance were reportedly used to successfully construct a retrieval framework and improve performance [41]. To further cope with large-scale complex retrieval problems in remote sensing, a two-steps strategy was reportedly used to obtain multi-hash codes, achieving a high retrieval accuracy over a short period of time [42]. Besides, a novel multi-label method based on fully convolutional network [67] is used for CBRSIR task, which shows great advantages over some single-label methods for interpreting complex remote sensing images.…”

Deep Hashing Based on Class-Discriminated Neighborhood Embedding

“…Despite their effectiveness, conventional hashing methods tend to require rather lengthy binary codes to achieve accurate CBIR results when dealing with complex multi-dimensional optical data, which generally makes other alternatives more convenient to effectively retrieve RS data [20], [21], [41], [42]. Among all the conducted research, deep-hashing models have recently exhibited great potential in RS due to the prominent success of CNNs to uncover highly discriminating features from aerial scenes [28], [43].…”