Work

Hybrid Approaches for Pneumonia Detection in X-rays: Combining CNNs and ML Classifiers

Year: 2025

Type: conference-paper

Abstract: Pneumonia is a disease that impacts millions of people worldwide, and X-ray image detection is one of the primary diagnostic tools used.This study presents a hybrid diagnostic approach combining Convolutional Neural Networks (CNNs) with traditional classifiers, namely Random Forest (RF) and Support Vector Machine (SVM), to detect pneumonia from chest X-ray images.Features were extracted from MobileNetV2, VGG16, and EfficientNetB0 and used to train RF and SVM models with hyperparameter tuning via GridSearchCV.Ensemble models were also explored: (i) CNNs + RF, (ii) CNNs + SVM, and (iii) RF + SVM.Experiments were conducted using a public pediatric dataset (5.856 X-rays) with stratified k-fold cross-validation and data augmentation.CNNs + RF achieved the highest 0.977 AUC and 91.9% accuracy, while individual models like VGG16 showed competitive performance (91.8% accuracy, 0.969 AUC).Results were statistically validated and showed strong potential for clinical support, particularly in settings with limited resources.In future work, we propose extending the approach to multiclass classification and refining model optimization strategies.

Source: Annals of Computer Science and Information Systems

DOI: 10.15439/2025f4310

Authors: Gabriele S. Araújo, Omar Andres C. Cortes, Olaf Reinhold, Fabio M. F. Lobato

Institutions: Universidade Estadual do Maranhão, Instituto Federal do Maranhão, Azerbaijan Cooperation University, The Co-operative University of Kenya, Universidade Federal do Oeste do Pará

Language: English

Topic: COVID-19 diagnosis using AI

Subfield: Radiology, Nuclear Medicine and Imaging

Field: Medicine

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: diamond

Funder: Conselho Nacional de Desenvolvimento Científico e Tecnológico