This document discusses feature selection and feature reduction. Feature selection is a process that chooses an optimal subset of features according to an objective function to reduce dimensionality, remove noise, and improve mining performance, speed of learning, predictive accuracy, and simplicity. Feature reduction transforms all original features into linear combinations, while feature selection only uses a subset of the original features. The document also lists some application areas of texture analysis, including food processing, medical image analysis, biometrics analysis, and global information systems.

1. FEATURE SELECTION AND
FEATURE REDUCTION

2. 3/6/2024 Department of Biomedical Engineering, SRMIST, KTR 2
IRIS SCAN IMAGE

3. 3/6/2024 Department of Biomedical Engineering, SRMIST, KTR 3
IRIS SYTEM
Pre
processing
Feature-
extraction
Identification
Verification
Stored
templates
Uniform
distribution
Reject
Accept
Iris scan image
capture
Iris
localization
Iris code comparison

4. Feature Selection
• Definition
• A process that chooses an optimal subset of features
according to a objective function
• Objectives
• To reduce dimensionality and remove noise
• To improve mining performance
• Speed of learning
• Predictive accuracy
• Simplicity and comprehensibility of mined results
4
3/6/2024 Department of Biomedical Engineering, SRMIST, KTR

5. Feature Reduction
vs.
Feature Selection
• Feature reduction
• All original features are used
• The transformed features are linear combinations of the
original features
• Feature selection
• Only a subset of the original features are selected
• Continuous versus discrete
5
3/6/2024 Department of Biomedical Engineering, SRMIST, KTR

6. Application Areas of Texture Analysis
3/6/2024 Department of Biomedical Engineering, SRMIST, KTR 6
Food processing
industry
Medical image
analysis
Biometrics analysis
(fingerprint, iris or retina,
etc.)
Global information system
(GIS) (for land, etc. analysis)