CDFIR: Cummulative distribution function based image retrieval
N Rajani, A Sreenivasa Murthy · 2017
Content Based Image Retrieval is a well-known retrieval process in the field of Image processing. CBIR is a special way of finding similar images from huge database. CBIR utilizes three rudimentary features like color, texture and shape which plays an essential role in image retrieval. The effective image retrieval process is the need and number of computations along with the rate of retrieval should be less and high respectively. Thus a simple function using cumulative distribution function is involved in the retrieval process. In this work, we propose a new method to determine similar images from large database with the use of shape feature. The Morphological processing is applied to an image to get shape feature i.e. boundary of the image. For the boundary extracted image a simple basic cumulative distribution function is applied and it results in similar intensity distribution for an image. The similarity measurement is performed using Euclidean distance. The retrieval process is compared with shape extracted feature, CDF applied feature and with edge detection algorithms. The outcome would be a less computation and good accuracy in finding the similar images.