Scribd
Upload
9 views

8 Feature Extraction

Uploaded by

Lingala Jayaprada
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
9 views

8 Feature Extraction

Uploaded by

Lingala Jayaprada
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 18

Feature Extraction

Overall Flow

Data Feature Object


Preprocessing Segmentation
Acquisition Extraction Recognition

Acquire data Improve Segment the Extract unique Assign a label to


in the form of quality of image into attributes of the an object based
images image regions/objects region/object on the attribute
Feature Extraction

There are two principal aspects of image feature extraction:


 Feature Detection: Feature detection refers to finding the features in an image, region,
or boundary.
 Feature Description: Feature description assigns quantitative attributes to the detected
features.
Feature Detection and Description
 Objective: To detect and describe information embedded in an image in other forms that are
more suitable than the image itself.
 Benefits:
- Easier to understand
- Require fewer memory, faster to be processed
- More “ready to be used”
 What kind of information we can use?
- Boundary, shape
- Region
- Texture
- Relation between regions
Features
 Features are categorized as local vs. global.
 A feature is local if it is applies to a member of a set,
 And global if it applies to the entire set,
where “member” and “set” are determined by the application.
Representation/Detection
➢ The ways of representing a region:
o External characteristics (its boundary): When the primary focus on shape characteristics
o Internal characteristics (its internal pixels): When the focus is on regional properties such as color and texture
o Whole image
➢ Methods used to represent the image/boundary/region:
o Boundary (Border) Following
o Chain Codes
o Boundary Approximations Using Minimum-Perimeter Polygons
o Signatures
o Boundary Segments
o Skeletons, Medial Axes, Distance Transforms
Feature Description

 The next step is to describe the region based on the chosen representation. By doing this, we quantify our
representation of the object
 Feature descriptors should be as insensitive as possible to variations in parameters such as scale,
translation, rotation, illumination, and viewpoint
Methods used for Feature Description
 Boundary Descriptors
o Geometrical descriptors : length, diameter, perimeter, eccentricity, curvature
o Shape Numbers
o Fourier Descriptors
o Statistical Moments
 Regional Descriptors
o Geometrical descriptors: Area, compactness, circularity ratio, Euler number
o Topological Descriptors
o Texture
o Moment Invariants
 Principal Components as Feature Descriptors
 Whole Image Features
o Harris-Stephens Corner Detector
o Maximally Stable Extremal Regions
 Scale-invariant Feature Transform (SIFT)
Chain codes
 The boundary is a good representation of an object shape and also requires a few
memory.
 Chain codes: represent an object boundary by a connected sequence of straight line
segments of specified length and direction.
 In order to represent a boundary, it is useful to compact the raw data (list of boundary pixels)
 Chain codes: list of segments with defined length and direction
▪ 4-directional chain codes
▪ 8-directional chain codes
Chain codes
It may be useful to down sample the data before computing the chain code
-to reduce the code dimension
-to remove small detail along the boundary
Chain codes
 Problem of a chain code: a chain code sequence depends on a starting point.
 Solution: treat a chain code as a circular sequence and redefine the starting point
so that the resulting sequence of numbers forms an integer of minimum
magnitude. The first difference of a chain code: counting the number of direction
change (in counterclockwise) between 2 adjacent elements of the code.
 E.g., 10103322 -> 3133030 (counting ccw)
and adding the last transition (circular sequence: 2 -> 1) -> 31330303
Boundary descriptors

 Simple descriptors
- length (e.g., for chain code: hor+vert+21/2*diagonal)
- diameter (length of the major axis)
- Eccentricity
- Curvature
Boundary descriptors

 Shape Numbers : Order of a shape: the number of digits


 Shape number: the first difference of smallest magnitude (treating the chain code as a circular sequence)
Boundary descriptors
 Fourier Descriptors
 The sequence of boundary points can be treated as a sequence of complex points in the complex plane
▪ It becomes a 1-D descriptor
▪ Can be DFT-transformed
Regional Descriptors
 Some simple descriptors
- Area A
- Perimeter P
- Compactness =perimeter2/area
- Circularity ratio Rc

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy