Open Access Open Access  Restricted Access Subscription Access

Images Search Based On Nearest Keyword’s Cluster In Database Of Multi-tagged Images

Mrunali Morkar, Chetna Ahirrao, Prerna More

Abstract


To create a better and robust system of searching the images from a huge dataset, we started working on a system which has an ability to provide optimal images from huge real-life databases. Our approach here is to maximize the accuracy of the displayed image by using a method, that we call ‘The NKS (Nearest Keyword Set) Method’. Here, the system will study the queries entered by the users and will ask for the tightest groups of points from the database satisfying the given keywords from the query. NKS method uses hash-based index structures which assists the system to achieve high scalability and speedup. Along with this, the system will have a ranking function which would rank the images having conflicts due to same values of NKS indexing.
Our experimental results on the real image datasets will show faster retrieval of a nearoptimal image as per users’ queries.


Full Text:

PDF

References


X. Cao, G. Cong, C. S. Jensen, and B. C. Ooi, “Collective spatial keyword querying,” in Proc. ACM SIGMOD Int. Conf. Manage. Data, 2011, pp. 373–384..

Songhe Feng, Zheyun Feng and Rong Jin,“Learning to Rank Image Tags With Limited Training Examples”, IEEE TRANSACTIONS ON IMAGE PROCESSING VOL: PP NO: 99 YEAR 2015

Vishwakarma Singh, Bo Zong, and Ambuj K. Singh,” Nearest Keyword Set Search in Multi-Dimensional Datasets”, IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, VOL. 28, NO. 3, MARCH 2016.


Refbacks

  • There are currently no refbacks.