Open Access Open Access  Restricted Access Subscription Access
Open Access Open Access Open Access  Restricted Access Restricted Access Subscription Access

Computation of Image Similarity with Time Series


Affiliations
1 Department of Computer Science and Engineering, Sri Vidya College of Engineering and Technology, Tamil Nadu, India
2 Department of Statistics, Manonmaniam Sundaranar University, Tamil Nadu, India
     

   Subscribe/Renew Journal


Searching for similar sequence in large database is an important task in temporal data mining. Similarity search is concerned with efficiently locating subsequences or whole sequences in large archives of sequences. It is useful in typical data mining applications and it can be easily extended to image retrieval. In this work, time series similarity analysis that involves dimensionality reduction and clustering is adapted on digital images to find similarity between them. The dimensionality reduced time series is represented as clusters by the use of K-Means clustering and the similarity distance between two images is found by finding the distance between the signatures of their clusters. To quantify the extent of similarity between two sequences, Earth Mover’s Distance (EMD) is used. From the experiments on different sets of images, it is found that this technique is well suited for measuring the subjective similarity between two images.

Keywords

Similarity Search, Vector Quantization, Similarity Measures, Clustering, EMD.
Subscription Login to verify subscription
User
Notifications
Font Size

Abstract Views: 237

PDF Views: 0




  • Computation of Image Similarity with Time Series

Abstract Views: 237  |  PDF Views: 0

Authors

V. Balamurugan
Department of Computer Science and Engineering, Sri Vidya College of Engineering and Technology, Tamil Nadu, India
K. Senthamarai Kannan
Department of Statistics, Manonmaniam Sundaranar University, Tamil Nadu, India
S. Selvakumar
Department of Statistics, Manonmaniam Sundaranar University, Tamil Nadu, India

Abstract


Searching for similar sequence in large database is an important task in temporal data mining. Similarity search is concerned with efficiently locating subsequences or whole sequences in large archives of sequences. It is useful in typical data mining applications and it can be easily extended to image retrieval. In this work, time series similarity analysis that involves dimensionality reduction and clustering is adapted on digital images to find similarity between them. The dimensionality reduced time series is represented as clusters by the use of K-Means clustering and the similarity distance between two images is found by finding the distance between the signatures of their clusters. To quantify the extent of similarity between two sequences, Earth Mover’s Distance (EMD) is used. From the experiments on different sets of images, it is found that this technique is well suited for measuring the subjective similarity between two images.

Keywords


Similarity Search, Vector Quantization, Similarity Measures, Clustering, EMD.