University of Turku, Dept. of Information
Technology
Spring 2016
Teacher: Jukka Teuhola
Contents: This course goes to the
fundamentals of data representation on bit level. The goal of data
compression is to reduce the size of the representation, with the
condition that the original data can be later recovered. Compression is
accomplished by eliminating part of the redundancy of the original
representation. The purposes of compression are to save storage space and to
reduce transmission time. The course first presents the basic information
theory needed to understand data compression. Thereafter, the most important source
coding (called also entropy coding) methods are presented, such
as Huffman and arithmetic coding. Modelling of the source data and its
redundancies is the key component in developing practical data compressors. Of
the source data types, emphasis is on the compression of symbol/number
sequences (mainly text data), but also lossless image compression
will be discussed.
Extent: 5 sp
Level: Advanced
Preliminary
knowledge:
Data structures and algorithms I, basics of probability calculus
Target audience: Students of computer science,
digital media, communication systems, or others, who are interested in the
theory and methods behind today's general-purpose compression software, such as
gzip, bzip2, JPEG, and others.
Course components: Starting lecture (2 h) on Wed
2.3.2016 at 8:15-10 in XXI (Agora), a few exercises, and an examination. The
course materials are available here:
- Lecture notes
- Powerpoint slides showing the key issues in
short
- Exercises
Examination
dates: 10.5,
13.6, and September 2016
Note: Usage of a pocket calculator is
allowed in the examination.