Lab Project

1
Lab Project

• Objective be familiar with JPEG
• Description
• Input an original 88 block gt original data
  (the value of each pixel is in the range of
  0,255) (given by the Instructor)
• Steps
• Shifted to the range of -128, 127
• FDCT (round 1.6-gt2, -1.6-gt-2, 1.5-gt2, 1.1-gt1))
  (PI 3.141592654)
• Quantized with the given quantization table (see
  the given example) (round)
• Zig-zag coded
• Converted to the intermediate symbol sequence
  with run-length encoding on AC coefficients
• Encoded to bit sequence with Huffman encoding
  (based on the given default Huffman codewords
  tables) gt compressed data
• The compression ratio
• original data size (i.e.
  648)/compressed data size
• Output print out all the results of the above
  steps into different files (The names and formats
  of output files should follow the files put on
  the web.)

2
Lab Project

• Team one or 2 students
• What should be turned in?
• Put all codes into one file (source_jpeg.c (.cpp)
  with clear comments (softcopy)
• Make sure TA can understand, compile, and run
  your codes on unix machines. All the codes should
  be in C or C.
• A report (2 lt page lt 5) with the source code
  file, and result files (Hardcopy)
• How will TA grade?
• Correctness TA will use his own data to test
  your codes.
• Clearness TA will check whether your code and
  report are understandable.
• Fairness
• Students in the same teams get the same scores.
• No copy! All students involved in copy will get
  ZERO.