A scalable approach for content based image retrieval in cloud datacenter

Abstract

The emergence of cloud datacenters enhances the capability of online data storage. Since massive data is stored in datacenters, it is necessary to effectively locate and access interest data in such a distributed system. However, traditional search techniques only allow users to search images over exact-match keywords through a centralized index. These techniques cannot satisfy the requirements of content based image retrieval (CBIR). In this paper, we propose a scalable image retrieval framework which can efficiently support content similarity search and semantic search in the distributed environment. Its key idea is to integrate image feature vectors into distributed hash tables (DHTs) by exploiting the property of locality sensitive hashing (LSH). Thus, images with similar content are most likely gathered into the same node without the knowledge of any global information. For searching semantically close images, the relevance feedback is adopted in our system to overcome the gap between low-level features and high-level features. We show that our approach yields high recall rate with good load balance and only requires a few number of hops.

Fusion feature for LSH-based image retrieval in a cloud datacenter

“…ICS adopts p-stable LSH to preserve fusion feature similarity and distributes the indexes to the Chord as evenly as possible. That is different from the traditional location approach [22], where DHTs access an image through the hash key of a single feature or key words.…”

“…In such a setting, it is the naïve solution that each feature is searched at a time, then the final results are given, after all the features is traversed. Obviously, since searching a feature needs a number of hops [22], multi-feature incurs high-cost communication. Analogously, it is hard to create a feature which takes into account all the feature such as color, texture and shape [7].…”

“…Precision is defined as the percentage of relevant images among the retrieved images. In order to compare with LRFIR [22], we have implemented the method and parameters which are set as described in [22]. Our aim is to evaluate the performance of the distributed indexes of a signal feature and multiple features.…”

“…However, the content-based image retrieval (CBIR) can find similar images through the image content instead of keywords. In our previous work, we present a CBIR system called LRFIR [22] based on a single feature for the distributed environment. It only adopts the texture feature to represent images, which is suitable for images containing texture.…”

Dynamic intelligence towards merging cloud and communication services

“…However, traditional search techniques only allow users to search images over exactmatch keywords through a centralized index. The paper by Liao et al (2014) entitled "A Scalable Approach for Content Based Image Retrieval in Cloud Datacenter" proposes a scalable image retrieval framework which can efficiently support content similarity search and semantic search in the distributed environment. Its key idea is to integrate image feature vectors into distributed hash tables (DHTs) by exploiting the property of locality sensitive hashing (LSH).…”