Lecture

Region Growing

It is a procedure that groups pixels or subregions into larger regions.

Region Growing is useful for:

  1. Segmenting images where regions of interests are not clearly defined by edges.
  2. Images has varying intensity levels that make traditional thresholding ineffective.

Images as Graphs

  • Graph Representation - Image can be modeled as a graph G=(V,E), where: V is a set of nodes representing pixels or pixel clusters. E is a set of edges indicating relationships between those nodes.

Graph Types:

  1. Undirected Graph
  2. Directed Graph

The types of graphs we are interested in are undirected graphs whose edges are further characterized by a matrix, W, whose element w(i, j) is a weight associated w/ the edge that connects nodes i and j.

Min Graph-Cut Segmentation Algorithm

  1. Specify a weighted graph, G = (V, E) in which V contains the points in the feature space, and E contains the edges of the graph.
  2. Add Source, and Sink Nodes.
  3. Assign Weights to Edges
  4. Compute the Minimum Cut
  5. Segment the graph and apply postprocessing if necessary.

Image Compression and Watermarking

Image Compression - reduces the amount of data required to represent a digital image.

Why is compression needed?

  • For storage and transmssion efficiency.
  • Compression can significantly reduce the size of an image. The effectiveness of compression depends on the content of the image.

NOTE: Data != Information Data and information are not synonymous. Data is the raw material used to convey information. Our goal is to minimize the amount of data while retaining as much information as possible. The same information can be represented using varying amounts of data. ex: “Hello, how are you doing today?” can be converted to “Hlo, hw r u dng tdy?”

Types of Data Redundancies

  1. Coding Redundancy:
  • each piece of information or event is assigned a sequence of code symbols called a code word.
  • Most 2-D intensity arrays use 8-bit codes to represent intensities, which may contain more bits than needed.
  1. Spatial and Temporal Redundancy
  • Spatial - Pixels in 2D intensity arrays are often similar to their neighbors, leading to unnecessary data replication
  • Temporal - In video sequences, similarities across frames can result in duplicated information.
  1. Irrelevant Information

Image Compression - Coding Redundancy

  1. Code - a list of symbols (letters, numbers, bits etc.)
  2. Code word - Sequence of symbols used to represent some information (ex: gray levels).
  3. Code word length - number of symbols in a code word.

Reading

Summary