Contact at mumbai.academics@gmail.com or 8097636691
Responsive Ads Here

Tuesday, 6 February 2018

EMR A Scalable Graph-based Ranking Model(2015)

EMR A Scalable Graph-based Ranking Model(2015)

INTRODUCTION

GRAPH-BASED ranking models have been deeply studied and widely applied in information retrieval area. In this paper, we focus on the problem of applying a novel and efficient graph-based model for content based image retrieval (CBIR), especially for out-of-sample retrieval on large scale databases. Traditional image retrieval systems are based on keyword search, such as Google and Yahoo image search. In these systems, a user keyword (query) is matched with the context around an image including the title, manual annotation, web document, etc. These systems don’t utilize information from images. However these systems suffer many problems, such as shortage of the text information and inconsistency of the meaning of the text and image. Content-based image retrieval is a considerable choice to overcome these difficulties. CBIR has drawn a great attention in the past two decades [1]–[3]. Different from traditional keyword search systems, CBIR systems utilize the low-level features, including global features (e.g., color moment, edge histogram, LBP [4]) and local features (e.g., SIFT [5]), automatically extracted from images. A great amount of researches have been performed for designing more informative low-level features to represent images, or better metrics (e.g., DPF [6]) to measure the perceptual
similarity, but their performance is restricted by many conditions and is sensitive to the data. Relevance feedback [7] is a useful tool for interactive CBIR. User’s high level perception is captured by dynamically updated weights based on the user’s feedback. Most traditional methods focus on the data features too much but they ignore the underlying structure information, which is of great importance for semantic discovery, especially when the label information is unknown. Many databases have underlying cluster or manifold structure. Under such circumstances, the assumption of label consistency is reasonable [8], [9]. It means that those nearby data points, or points belong to the same cluster or manifold, are very likely to share the same semantic label. This phenomenon is extremely important to explore the semantic  relevance when the label information is unknown. In our opinion, a good CBIR system should consider images’ lowlevel
features as well as the intrinsic structure of the image database. Manifold Ranking (MR) [9], [10], a famous graph-based ranking model, ranks data samples with respect to the intrinsic geometrical structure collectively revealed by a large number of data. It is exactly in line with our consideration. MR has been widely applied in many applications, and shown to have excellent performance and feasibility on a variety of data types, such as the text [11], image [12], [13], and video[14]. By taking the underlying structure into account, manifold ranking assigns each data sample a relative ranking score, instead of an absolute pair wise similarity as traditional ways. The score is treated as a similarity metric defined on the manifold, which is more  meaningful to capturing the semantic relevance degree. He et al. [12] firstly applied MR to CBIR, and significantly improved image retrieval performance compared with state-of-the-art algorithms. However, manifold ranking has its own drawbacks to handle large scale databases – it has expensive computational cost, both in graph construction and ranking computation stages. Particularly, it is unknown how to handle an out-of-sample query (a new sample) efficiently under the existing framework. It is unacceptable to re compute the model for a new query. That means, original manifold ranking is inadequate for a real world CBIR system, in which the user provided query is always an out-of-sample. In this paper, we extend the original manifold ranking and propose a novel framework named Efficient Manifold Ranking (EMR). We try to address the shortcomings of manifold ranking from two perspectives: the first is scalable graph construction; and the second is efficient computation, especially for out-of-sample retrieval. Specifically, we build an anchor graph on the database instead of the traditional k-nearest neighbor graph, and design a new form of adjacency matrix utilized to speed up the ranking computation. The model has two separate stages: an offline stage for building (or learning) the ranking model and an online stage for handling a new query. With EMR, we can handle a database with 1 million images and do the online retrieval in a short time. To the best of our knowledge, no previous manifold ranking based algorithm has run out-of-sample retrieval on a database in this scale. A preliminary version of this work previously appeared as [13]. In this paper, the new contributions are as follows: • We pay more attention to the out-of-sample retrieval (online stage) and propose an efficient approximate method to compute ranking scores for a new query in Section 4.5. As a result, we can run out-of sample retrieval on a large scale database in a short time. • We have optimized the EMR code1 and re-run all the experiments (Section 5). Three new databases including two large scale databases with about 1 millions samples are added for testing the efficiency of the proposed model. We offer more detailed analysis for experimental result. • We formally define the formulation of local weight estimation problem (Section 4.1.1) for building the anchor graph and two different methods are compared to determine which method is better (Section 5.2.2). The rest of this paper is organized as follows. In Section 2, we briefly discuss some related work and in Section 3, we review the algorithm of MR and make an analysis. The proposed approach EMR is described in Section 4. In Section 5, we present the experiment results on many real world image databases. Finally we provide a conclusions in Section 6.
IMPLEMENTATION
  • Admin
In this module, the Admin has to login by using valid user name and password. After login successful he can do some operations such as upload images, view uploaded images, view all data sets of images, list of all searching history,  view all image ranking and view all users, search images and logout.
Upload Images
In this module, the admin can upload n number of images. Admin want to upload new image then he has enter some fields like image name, image color, image description, image type, living place, browse the image file and upload. After uploading successfully he will get a response from the server. Initially new uploaded image rank is zero. After viewing that image rank will re-rank.
View data set of Images
In this module, the Admin can view the all type’s images available in server. If admin wants to view all type of images, then click on data set images button, it will give response to user with key words such as human, birds, animals, Insects, fruits, trees and non living objects.
Search History
This is controlled by admin; the admin can view the search history details. If he clicks on search history button, it will show the list of searched user details with their tags such as user name, user searched for image name, time and date.
Rank of images
In this module, the admin can view the list of ranking images. If admin click on list of ranking images, then the server will give response with their tags image and rank of image.

  • User
In this module, there are n numbers of users are present. User should register before doing some operations. And register user details are stored in user module.  After registration successful he has to login by using authorized user name and password. Login successful he will do some operations like view my details, search images, request secrete key and logout. The user click on my details button then the server will give response to the user with all details such as user name, phone no, address, e mail ID and location. Before searching any images user should request a secrete key to admin, then the admin will generate a secrete key for particular user and send to the user. After getting a secrete key user can search the images base on query and field like image name, image color, image usage and image type. And server will give response to the user, then that image rank will be increased.

No comments:

Post a Comment