C Java Tutor 客服在線

代做國外C C++ Java程序 QQ: 1067665373 Email:cjtutor@foxmail.com

« Monopoly Game in Cackdoor Web Server »

Image Retrieval System

 

 Implement a Simple Content-Based Image Retrieval System

 

Description:

  CSS 490 B Multimedia Data Processing

This project is to implment a simple Content-Based Image Retrieval system based on two different color histogram comparison methods.

 

1.     Test Image Database

This test image database includes 100 true-color images in .jpg format.

 

2.     Color Histogram

Color histogram comparison is a simple but effective apporach in CBIR systems. Here are two ways to combine the information from 3 color channels (R, G, B):

 

A.    Intensity Method

                                                                     (1)

By this way, the 24-bit of RGB (8 bits for each color channel) color intensities can be transformed into a single 8-bit value. The histogram bins selected for this case are listed below:

 

H1:; H2:H3:;

H4:      ……      ……      ……      H6:;

… … …

H25:;

 

B.    Color-Code Method

The 24-bit of RGB color intensities can be transformed into a 6-bit color code, composed from the most significant 2 bits of each of the three color components, as illustrated in the following figure.

 

       The 6-bit color code will provide 64 histogram bins.

For example, the R, G, and B values for a pixel are 128, 0, and 255 respectively. So the bit representations of them are 10000000, 00000000, and 11111111. Then the 6-bit color code value will be 100011.

In color code, there will be 64 bins with H1: 000000, H2: 000001, H3: 000010, … H64: 111111

 

3.     Histogram Comparison

You need to implement the distance metrics for histogram comparison. Let Hi(j) denote the number of pixels in jth bin for the ith image. Then the difference between the ith image and kth image can be given by the following distance metric:

¨     

                                                                           (2)

where Mi*Ni is the number of pixels in image i, and Mk*Nk is the number of pixels in image k.

 

 

Preliminary demo (in groups):

Show the graphic user interface is up running. It should allow users to browse the image database, select the query image, and to view the retrieved images. At this stage, it is not required that the retrieval algorithm is properly implemented so the retrieved images can be any images from the database.

 

 

 

  • 相關文章:

發表評論:

◎歡迎參與討論,請在這里發表您的看法、交流您的觀點。

最新評論及回復

最近發表

Powered By

Copyright 代寫C.

在線客服

售前咨詢
售后咨詢
微信號
Essay_Cheery
微信
悉尼assignment代写,北美作业代写,代写毕业论文-100%原创 北美代写,Homework代写,Essay代寫-准时✔️高质✔最【靠谱】 墨尔本assignment代写,代写毕业论文,paper代写-51作业君 北美代写,程序代做,程序代写,java代写,python代写,c++代写,c代写 英国代写paper,python代写,Report代写,编程代写-程序代写网 北美代写essay,程序代写,Java代写代做,Java代考-焦点论文 澳大利亚essay代写,编程代写,代码代写,程序代写-三洋编程 加拿大essay代写|程序代写代做||Python代写|Matlab代写-Meeloun 澳大利亚代写,代写essay,代写毕业论文,留学生代写-小马代写 日本代写,北美作业代写,新加坡代写,essay代写-无时差服务