A New Approach for Retrieval of Image with Clustering Techniques and Local Mean Histogram

R. Tamilkodi, G. Rosline Nesa Kumari and S. Maruthuperumal

Characteristic taking out and identical is at the support of several computer revelation harms such as object identification or formation movement. The algorithms and the relevant images used surrounded by this work are alienated into two categories: color and texture. By combining the properties of two features one can get a better retrieval performance than using single method alone. This article proposed a new retrieval system by combining the properties of color and texture we can name the method as clustering techniques and local mean in short one can call it as CTALM. This method works on the principle of extraction the clusters from the query image provided. Based on the clusters one can find the local mean of the cluster image created. Once the local mean image has been generated find the average mean and store. The stored value is compared with the database image and similar assessment images are retrieved. The output image generated is seems to be similar to the query image supplied. Proposed method CTALM is compared with other methods and has been identified that this method shows good accuracy in retrieval.

Volume 11 | Issue 6

Pages: 62-71