DeepFusion-CNN: A Novel Context-awareNetwork for ECG-based Gender Classification

Abstract

Electrocardiogram (ECG)-based gender identification, which utilizes the electrical activity of the heart, hasemerged as a promising approach in biometric and healthcare applications. This study introduces DeepFusion-CNN, acontext-aware fusion framework that integrates VGG-19, DenseNet-121, and ResNet-152 using a validation-drivenadaptive weighting strategy to improve gender classification performance. Unlike conventional ensemble approachesthat use a static averaging strategy, the proposed approach adaptively adjusts each sub-model's contribution based on itsvalidation performance, enabling improved feature representation and classification robustness. This adaptive fusionmechanism allows better-performing models to contribute more significantly, leading to improved overall predictionaccuracy compared to individual models and static fusion strategies. The ECG signals are preprocessed using band-passfiltering, followed by R-peak identification using the Pan–Tompkins algorithm. The processed signals are thensegmented and converted into 225×225×3 two-dimensional images, making them suitable for transfer learning with pretrained convolutional models. To maintain a fair evaluation, the data is partitioned on a subject basis before anyaugmentation, and augmentation is restricted to the training portion only. The framework is evaluated on the PTB andCYBHi datasets, achieving accuracies of 99.08% and 99.13%, respectively. Ablation test results indicate that thefeature quality and classification performance are improved after preprocessing and the context-aware fusion strategy.The proposed framework shows strong potential for ECG-based gender classification and could serve as a usefulfoundation for future advancements in biometric systems and healthcare applications.