Investigating the properties of Zernike moments for robust content based image retrieval

The issues of efficient search and retrieval of relevant images from large image database is becoming more critical nowadays as there are a huge amount of digital images particularly in the internet. This paper presents shape-based image retrieval using Zernike Moments method. Experimental investigations have been carried out using MPEG-7 shape image database to test Zernike Moments properties and its robustness to rotation, translation, scale and the effect of noise. The findings of the experimental investigations shows that Zernike Moments properties are invariant to rotation, and translation with low sensitivity to noise, however is not invariant to scale.