6.5 Analysis
About1. Introduction2. Overview3. GUI4. Image Signatures5. Unsupervised Filters6. Results & Analysis7. BioFilters8. NeuralFilters9. Duplicated Documents10. Face Recognition11. Auto Part Recognition12. Dynamic Library13. NeuralNet Filter14. Segment Variation15. TV Advertisements16. Counting & Tracking17. Image PreProcessing18. Image Processing19. Batch Job20. Parameters21. Input Option22. Application Developers23. Reference Manual24. Support Services25. Readme.txt
6.1 Wheel Rim 
6.2 N:N Matching 
6.3 Three Output Files 
6.4 Checking Results 
6.5 Analysis 
6.6 False Rejection 
6.7 Threshold 
[Home][6. Results & Analysis][6.5 Analysis]

 

6.5    Analysis

Analysis means to calculate 3 indexes and a composite index. We will first  introduce several variables.

Possible Matches

    Let the total number of images be N, the Possible Matches will be N*N. In our example:

    N * N = 100 * 100 = 10,000.

Attrasoft Matches

    The number of retrieved matches is listed in the last line of b1.txt. Go to the end of b1.txt, you will see something like this:

    Total Number of Matches = 812

Actual Match

    This number depends on your problem. It should be the first number in b1_matchlist.txt. In our example, it is 400.

Attrasoft Found Pairs

    Click “Unsupervised/N:M Analysis/Check (b1_matchlist.txt required)” menu item to get this number, as discussed in the last section.

    Now that you have all of the numbers, you can make an analysis. The analysis consists of three individual rates and a Composite Index:

Positive Identification Rate

    Positive Identification Rate = the results of clicking “Unsupervised/N:M Analysis/Check (b1_matchlist.txt required)” menu item divided by the first number in file, b1_matchlist.txt.  

    In our example, it is 218/400. We will show you how to get this number to 100% in the next chapter.

Elimination Rate

    The Elimination Rate is 1 minus the number at the end of b1.txt divided by the number of possible matches. This number should be normalized so that if all mismatches are eliminated, it should be 100%. In our example, this number is:

    Elimination Rate = 1 – 812/10,000

    Normalized Elimination Rate = 1 –  (812 – 218) / (10,000 – 400)

Hit Ratio

    The Hit Ratio is the number indicated by “BioFilter/Check (b1_matchlist.txt required)” menu item divided by the number at the end of b1.txt. In our example, this number is: 218/812.

Composite Index

    Finally, an identification is measured by multiplication of Positive Identification Rate * Elimination Rate * Hit Ratio. In our example, this number is (218/400) * (1 – 812/10000) * (218/812).

 

The Unsupervised Filter is not accurate. It is used to assess how hard your image recognition problem is.

 

[Home][About][1. Introduction][2. Overview][3. GUI][4. Image Signatures][5. Unsupervised Filters][6. Results & Analysis][7. BioFilters][8. NeuralFilters][9. Duplicated Documents][10. Face Recognition][11. Auto Part Recognition][12. Dynamic Library][13. NeuralNet Filter][14. Segment Variation][15. TV Advertisements][16. Counting & Tracking][17. Image PreProcessing][18. Image Processing][19. Batch Job][20. Parameters][21. Input Option][22. Application Developers][23. Reference Manual][24. Support Services][25. Readme.txt]

Copyright (c) 1998 - 2006 Attrasoft, Inc. All rights reserved.

gina@attrasoft.com