The following files are used to test performance:
---------------------------------------------
File_name
File_size
Percentage_to_image
---------------------------------------------
1.data
36,787
3%
2.data
61,830
4%
3.data
76,347
5%
4.data
138,822
10%
5.data
272,106
19%
6.data
329,330
23%
7.data
512,174
36%
8.data
600,054
42%
9.data
608,070
42%
10.data
753,534
52%
11.data
848,590
59%
12.data
1,132,338
79%
13.data
1,139,291
79%
14.data
1,146,243
80%
15.data
1,404,445
98%
----------------------------------------------
Conclusions:
1) When file size is just more than 50% of image size, the quality of encrypted image
is still satisfying. And it is still difficult to find the difference between
new image and reference image ( base image).
2) When file size is more than 70%, you still can find the outline of the original
image, but you can find the image quality is becoming worse and worse.
3) When the size of file to be encoded in image is more than 95% of original image,
you can see some of the outline of image, but in some area, the original pixels
are totally overlapped.
4) considering other technologies such as data compress, data encryption, and image
preprocessing such as image smoothing, we should be able to get better results.
Ba-jian, 2009, 1, 9
---------------------
P.S. Sina blog doesn't support BMP files, so images are not uploaded.
However you can try it using the source code.
---------------------------------------------
File_name
---------------------------------------------
Conclusions:
Ba-jian, 2009, 1, 9
---------------------
P.S. Sina blog doesn't support BMP files, so images are not uploaded.
However you can try it using the source code.