Spherical hashing

Heo, Jae-Pil; Lee, Youngwoon; He, Junfeng; Chang, Shih‐Fu; Yoon, Sung‐Eui

doi:10.1109/cvpr.2012.6248024

Cited by 110 publications

(7 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Figure 2 shows the Precision Recall (PR) curve of the hash code on the CIFAR-10 and NUS-WIDE datasets. In order to evaluate the effectiveness of the proposed method, we compare the results of the proposed method and the comparison method, including DDH [10], UH-BDNN [4], ITQ [15], AGH [6], SpH [16], PCAH [17], LSH [18]. The length of the hash code in the CIFAR-10 experiment is 64 bits, and the NUS-WIDE experiment is 32 bits.…”

Section: Performance Analysismentioning

confidence: 99%

Unsupervised Deep Embedded Hashing for Large-Scale Image Retrieval

Wang

2021

IEICE Trans. Fundamentals

View full text Add to dashboard Cite

Hashing methods have proven to be effective algorithm for image retrieval. However, learning discriminative hash codes is challenging for unsupervised models. In this paper, we propose a novel distinguishable image retrieval framework, named Unsupervised Deep Embedded Hashing (UDEH), to recursively learn discriminative clustering through soft clustering models and generate highly similar binary codes. We reduce the data dimension by auto-encoder and apply binary constraint loss to reduce quantization error. UDEH can be jointly optimized by standard stochastic gradient descent (SGD) in the embedd layer. We conducted a comprehensive experiment on two popular datasets.

show abstract

Section: Performance Analysismentioning

confidence: 99%

Unsupervised Deep Embedded Hashing for Large-Scale Image Retrieval

Wang

2021

IEICE Trans. Fundamentals

View full text Add to dashboard Cite

show abstract

“…All aforementioned unimodal hashing models cannot generate balanced code matrix. Spherical hashing (SpH) [9] and global hashing system (GHS) [23] quantize the distance between a data point and a special point. The closer half to a special point is denoted as 1 while the further half is denoted as 0.…”

Section: A Unimodal Hashingmentioning

confidence: 99%

Learning Decorrelated Hashing Codes for Multimodal Retrieval

Tian

2018

Preprint

View full text Add to dashboard Cite

In social networks, heterogeneous multimedia data correlate to each other, such as videos and their corresponding tags in YouTube and image-text pairs in Facebook. Nearest neighbor retrieval across multiple modalities on large data sets becomes a hot yet challenging problem. Hashing is expected to be an efficient solution, since it represents data as binary codes. As the bit-wise XOR operations can be fast handled, the retrieval time is greatly reduced. Few existing multimodal hashing methods consider the correlation among hashing bits. The correlation has negative impact on hashing codes. When the hashing code length becomes longer, the retrieval performance improvement becomes slower. In this paper, we propose a minimum correlation regularization (MCR) for multimodal hashing. First, the sigmoid function is used to embed the data matrices. Then, the MCR is applied on the output of sigmoid function. As the output of sigmoid function approximates a binary code matrix, the proposed MCR can efficiently decorrelate the hashing codes. Experiments show the superiority of the proposed method becomes greater as the code length increases.

show abstract

“…The final step of SpH [40] is the same as our method, so SpH also generates a balanced code matrix. However, SpH searches the locations of special points in the entire space, which makes it difficult to find a good solution.…”

Section: Spherical Hashing (Sph)mentioning

confidence: 99%

“…We evaluate our methods by comparing to seven hashing methods which includes: Iterative Quantization (ITQ) [16], Isotropic Hashing (IsoH) [17], Harmonious Hashing (HH) [36], Spectral Hashing (SH) [13], Inductive Manifold Hashing (IMH) [12], Orthogonal K-means (ok-means) [43] and Spherical Hashing (SpH) [40]. Our data-dependent and data-independent are denoted as GHS-DD and GHS-DI, respectively.…”

Section: A Protocols and Baselinesmentioning

confidence: 99%

Global Hashing System for Fast Image Search

Tian

Tao

2017

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

Abstract-Hashing methods have been widely investigated for fast approximate nearest neighbor searching in large datasets. Most existing methods use binary vectors in lower dimensional spaces to represent data points, which are usually real vectors of higher dimensionality. However, according to Shannon's Source Coding Theorem (SSCT) in information theory, it is logical to represent low-dimensional real vectors with high-dimensional binary vectors, since a binary bit contains less information than a real number. We design a novel hashing method based on this principle. Data points are first embedded in a low-dimensional space, and then the Global Positioning System (GPS) method is introduced but modified for hashing. We devise data-independent and data-dependent methods to distribute the "satellites" at appropriate locations. Benefitting from the rationale of SSCT and rules on distributing satellites in a GPS, our data-dependent method outperforms other methods in different-sized datasets from 100K to 10M. By incorporating the orthogonality of the code matrix, both our data-independent and data-dependent methods are particularly impressive in experiments on longer bits.

show abstract

Spherical hashing

Cited by 110 publications

References 22 publications

Unsupervised Deep Embedded Hashing for Large-Scale Image Retrieval

Unsupervised Deep Embedded Hashing for Large-Scale Image Retrieval

Learning Decorrelated Hashing Codes for Multimodal Retrieval

Global Hashing System for Fast Image Search

Contact Info

Product

Resources

About