.png){kind=icon}
.png){kind=icon}
Return
Distortion Minimization Hashing
.png){kind=icon}
OA
AI
DOI：10.1109/ACCESS.2017.2763600.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Application of the hashing method to large-scale image retrieval has drawn much attention because of the high efficiency and favorable accuracy of the method. Its related research generally involves two basic problems: similarity-preserving projection and information-preserving quantization. Most previous works focused on learning projection approaches, while the importance of quantization strategies was ignored. Although several hashing quantization models have been recently proposed to improve retrieval performance by assigning multiple bits to projected directions, these models still suffer from suboptimal results, as the critical information loss that occurs in the quantization procedure is not considered. In this paper, to construct an effective quantization model, we utilize rate-distortion theory in the hashing quantization procedure and minimize the distortion to reduce the information loss. Furthermore, combining principal component analysis with our quantization strategy, we present a quantization-based hashing method named distortion minimization hashing. Extensive experiments involving one synthetic data set and three image data sets demonstrate the superior performance of our proposed methods over several quantization techniques and state-of-the-art hashing methods.
Keywords:
Machine learning
image retrieval
hashing quantization
rate-distortion theory
iterative optimization
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
3.6
Papers:
9.8W
Citations:
29.4W