IRMA-International.org: Creator of Knowledge
Information Resources Management Association
Advancing the Concepts & Practices of Information Resources Management in Modern Organizations

Feature Extraction and Feature Selection Procedures for Medical Image Analysis

Feature Extraction and Feature Selection Procedures for Medical Image Analysis
View Sample PDF
Author(s): Soumya Gupta (University School of Automation and Robotics, Guru Gobind Singh Indraprastha University, India)and Sia Gupta (University School of Automation and Robotics, Guru Gobind Singh Indraprastha University, India)
Copyright: 2025
Pages: 60
Source title: Computer-Assisted Analysis for Digital Medicinal Imagery
Source Author(s)/Editor(s): Amit Sinha (ABES Engineering College, Ghaziabad, India), Pranshu Saxena (School of Computer Science Engineering and Technology, Bennett University, Greater Noida, India), Sanjay Kumar Singh (University School of Automation and Robotics, Guru Gobind Singh Indraprastha University, East Delhi, India)and Harikesh Singh (JSS Academy of Technical Education, Noida, India)
DOI: 10.4018/979-8-3693-5226-7.ch010

Purchase

View Feature Extraction and Feature Selection Procedures for Medical Image Analysis on the publisher's website for pricing and purchasing information.

Abstract

This chapter delves into the critical processes of feature extraction and selection in medical image analysis, essential for transforming raw data into actionable insights. It begins with preprocessing techniques, including noise reduction using linear and nonlinear filters, to enhance image quality. Intensity-based methods utilize pixel statistics, while texture analysis techniques like Local Binary Patterns, co-occurrence matrices, wavelets, Fourier transforms, and orientation histograms capture intricate patterns. Deep learning-based features, especially autoencoders, automatically learn hierarchical data representations. For feature selection, filter methods evaluate relevance independently, wrapper methods iteratively train models to identify optimal subsets, and embedded methods integrate selection within training, promoting sparsity. Dimensionality reduction techniques like Principal Component Analysis (PCA) condense feature spaces, retaining essential information.

Related Content

Rubina Ghani. © 2026. 14 pages.
Amanullah Khan, Mohsin Qadeer, Salman Shareef. © 2026. 20 pages.
Amanullah Khan, Salman Shareef, Mehak Hafiz. © 2026. 18 pages.
Soobia Saeed, Mohsin Qadeer. © 2026. 14 pages.
Amanullah Khan, Salman A. Shareef. © 2026. 18 pages.
Muhammad Zubair, Bushra Khalid. © 2026. 30 pages.
Muhammad Bilal Ashraf. © 2026. 12 pages.
Body Bottom