Tasks

1. (4 pts) Write a well-documented program or find existing software to count the frequency of each pixel value in an ".img" image file.  Run this program on sinan.img and earth.img
• (submit) Using Microsoft Excel (or any other histogram-making software), produce histograms of the computed frequencies for each image. Your sinan histogram should look like the histogram in the textbook's Figure 11.2.
• (submit) a copy of your program if you wrote one, else indicate which existing software you used.
• (submit) Indicate the exact number of pixels in each orginal image having the values 0,1,2,3 and 127,128,129 and 252,253,254,255
2. (4 pts) Differential encoding
   • Create an "8-bit differential encoding" of sinan.img (and earth.img) named sinan.diff.img (and earth.diff.img) using the program you wrote in Project 1.
   • (submit) Use the program from the previous question to create a histogram for the frequency of difference values 0-255 for sinan.diff.img (and earth.diff.img). Your sinan histogram should contain the same info as contained in Figure 11.3. The two histograms will appear different since Figure 11.3's x-axis goes from -255 to +255, while your histogram's x-axis goes from 0-255. Remember that a difference of -255 is the same as a difference of +1.
   • (submit) Indicate the exact number of pixel differences your historgrams have for differences of 0, 1, and 255
3. (3 pts) (submit) Use xv, GIMP or other software to view and print copies of sinan.img and sinan.diff.img (and earth.img and earth.diff.img.) To view an ".img" image, use Sayood's program "convpgm.c" to produce a well-known format ".pgm" file, and then view/print the ".pgm" file. Take some time to understand why the difference images look like they do.
4. (4 pts Extra Credit for all students) Use Sayood's huff_enc program to compute a huffman code for both sinan.img and sinan.diff.img. Save the two code files separately. huff_enc generates a fixed length code file of 1280 bytes. The format of the code file is 256 4-byte "codes" followed by 256 1-byte "lengths". The first length tells how many least significant bits of the first 4-byte code are the Huffman code for a pixel value (or pixel difference value for a difference image) of 0. The second length tells how many least significant bits of the second 4-byte code are the Huffman code for the pixel value of 1, and so on. What are the Huffman codes for both sinan.img and sinan.diff.img for the pixel values and pixel difference values, respectively, of: 126,127,128,129,130 and 253,254,255,0,1,2,3,4,5,6? Looking at the two code files, argue whether the generated Huffman codes are consistent with the histograms you produced in parts 1 and 2 above.

Notes

1. Students may work individually or in groups of 2.   Students in different groups may NOT compare answers prior to submitting their work.  See the syllabus for a full statement of the lateness and academic honesty policies.
2. All source programs must be well documented.