A Hybrid Feature Extraction and Machine Learning Framework for Multiclass Medical Image Classification

In this paper, we present a systematic hybrid handcrafted feature extraction framework for medical image classification based on the integration of frequency-domain, statistical, texture and distribution-based descriptors in a single architecture. First, grayscale images are preprocessed to fixed resolution 128 × 128 pixels and normalized in the range [0, 1]. A two-dimensional DCT is subsequently employed to obtain frequency-domain coefficients that are vectorized and decomposed based on Benford’s law in order to encapsulate statistical characteristics found within the frequency plane. Simultaneously, compact global intensity and texture characteristics are obtained by first-order statistical features, Local Binary Pattern (LBP) descriptors, Gray-Level Co-occurrence Matrix (GLCM) features and normalized intensity histograms. Andrews curve embedding is applied to the Benford, statistical, LBP, and GLCM descriptors to represent higher-order feature interactions. All extracted descriptors are then stacked together as a compact feature vector, which is processed using feature-scaling and mutual-information based-feature selection techniques. An Extreme Gradient Boosting (XGBoost) model then classifies the resulting representation. A representation produced using an Extreme Gradient Boosting (XGBoost) model is then used to classify the exposure. The introduced approach provides an effective, simple to interpret and discriminative representation that can be directly used for accurate disease classification from medical images by designing image representations leveraging intermodal complementary image attributes. Experimental results demonstrated that the proposed multiclass COVID–Normal–Pneumonia chest X-ray classification framework achieved 93.74% accuracy, 96.07% macro precision, 86.40% macro sensitivity, 98.83% macro specificity, and 90.51% macro F1-score on COVID-19 Radiography Database. Comparative analysis confirmed competitive performance with improved specificity and reliable multiclass disease discrimination capability.

Authors

Publication Details

Journal
The International Journal of Maritime Engineering
Published
2026-11-13
DOI
https://doi.org/10.5750/ijme.v168ia6(s).2033
Primary Topic
COVID-19 diagnosis using AI
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

A Hybrid Feature Extraction and Machine Learning Framework for Multiclass Medical Image Classification

Kalli Srinivasa Nageswara Prasad, Prashanthi Peram, Attili Venkata Ramana, P. Swetha et al.
1 links
The International Journal of Maritime Engineering
COVID-19 diagnosis using AI
article

A Hybrid Feature Extraction and Machine Learning Framework for Multiclass Medical Image Classification

Kalli Srinivasa Nageswara Prasad, Prashanthi Peram, Attili Venkata Ramana, P. Swetha, V. Sravan Kiran
article en

Abstract

In this paper, we present a systematic hybrid handcrafted feature extraction framework for medical image classification based on the integration of frequency-domain, statistical, texture and distribution-based descriptors in a single architecture. First, grayscale images are preprocessed to fixed resolution 128 × 128 pixels and normalized in the range [0, 1]. A two-dimensional DCT is subsequently employed to obtain frequency-domain coefficients that are vectorized and decomposed based on Benford’s law in order to encapsulate statistical characteristics found within the frequency plane. Simultaneously, compact global intensity and texture characteristics are obtained by first-order statistical features, Local Binary Pattern (LBP) descriptors, Gray-Level Co-occurrence Matrix (GLCM) features and normalized intensity histograms. Andrews curve embedding is applied to the Benford, statistical, LBP, and GLCM descriptors to represent higher-order feature interactions. All extracted descriptors are then stacked together as a compact feature vector, which is processed using feature-scaling and mutual-information based-feature selection techniques. An Extreme Gradient Boosting (XGBoost) model then classifies the resulting representation. A representation produced using an Extreme Gradient Boosting (XGBoost) model is then used to classify the exposure. The introduced approach provides an effective, simple to interpret and discriminative representation that can be directly used for accurate disease classification from medical images by designing image representations leveraging intermodal complementary image attributes. Experimental results demonstrated that the proposed multiclass COVID–Normal–Pneumonia chest X-ray classification framework achieved 93.74% accuracy, 96.07% macro precision, 86.40% macro sensitivity, 98.83% macro specificity, and 90.51% macro F1-score on COVID-19 Radiography Database. Comparative analysis confirmed competitive performance with improved specificity and reliable multiclass disease discrimination capability.

The International Journal of Maritime EngineeringVol. 168(A6(S))
Good health and well-being
COVID-19 diagnosis using AI
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon