Binary Plane Technique Based Color Quantization for Content Based Image Retrieval

T Esther Ratna,N Subash Chandra

doi:10.14419/ijet.v7i3.1.16814

Abstract

Extracting accurate informative file from a high volume of graphic files is a challenging task. This paper focus on presenting a new color indexing approach using the histogram features. Two histogram features like maximum color histogram and minimum color histogram are computed and are vector quantized to constitute a feature vector. Bit plane technique is used to map these features based upon it value at the respective position. The ultimate goal of any retrieval method is to attain higher precision within a short span of time that could be achieved if the data is in compressed to accomplish this the image is compressed using binary plane technique. The result analysis depicts the performance of the proposed approach under lossy and lossless modes and found that when operated in lossy it attain effective precision rate in a speculated amount of time.

Highlights

Retrieving the data based on text keywords as query is what it is observed in many of the search engines and digital libraries but it involves many tasks in retrieving the image data
Content based image retrieval (CBIR) is an approach to extract the relevant images from the digital library using visual features like color, shape and texture
The work consists of two new color features named as color histogram feature, bin pattern histogram feature which are directed to vector quantization to form color quantizer and indexed bit map image

Summary

Introduction

Retrieving the data based on text keywords as query is what it is observed in many of the search engines and digital libraries but it involves many tasks in retrieving the image data. Content based image retrieval (CBIR) is an approach to extract the relevant images from the digital library using visual features like color, shape and texture. Most of the standard compression techniques like JPEG does not consider the retrieval process [3], there are some methods like vector Quantization (VQ) which involve this despite of suitable power they are observed to be slow for compression [4]. In simpler terms, it can be termed as the problem of showing an image with high efficiency in binary form.

Related Work

Pseudo Code for Retrieval Process

Experimental Results