Content Based Retrieval:
Two Phase Search
- Background Information
- Relational Database Idea
- Multi-Attribute Selection with Single Attribute Indices
- We can guarantee not missing out on any object
- Example of Spatial Join: Map of countries and rivers
- Shape Matching Example
- General Rectangular Shape Covers
- Additive Rectangular Shape Covers
- Sequential Descriptions
- The first rectangles added are the ones that best
show most important features
- Uniqueness of each description
- How do we store these rectangles in the database?
- The position (x,y) of the rectangle is the center
of it
- Subsequent height and length
- From here we can implement any type of indexing system
- Examples of Matching Algorithms
- Full Match
-has the correct shape and position
- Match With Shift
-doesn't matter if the shape is shifted or not
- Match With Uniform Scaling
-The actual size of the object doesn't matter
- Match With Independant Scaling
-the axes can possibly have non-uniform scaling
-used when picture is taken at angle instaed of straight on