Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
March 6, 2026Journal of ImagingOpen Access

Optimizing Radiographic Diagnosis Through Signal-Balanced Convolutional Models

View Full Paper
Ask AI
Bookmark
Share

Authors

SNSakina Juzar NeemuchwalaRARashid AliQAQamar Abbas

Discussion

Loading...

Member takes

Overview

Demonstrates improved diagnostic accuracy in pulmonary disorders using a signal-aware framework.

Key Points

  • The aim is to enhance the reliability and transparency of chest radiograph interpretations using deep learning.
  • Utilized the COVID-19 Radiography Dataset consisting of 21,165 chest X-ray images.
  • Trained baseline CNN, ResNet-50, and EfficientNetB3 models under various configurations.
  • Applied signal fidelity analysis for quantitative evaluation using Structural Similarity Index Measure.
  • Implemented Gradient-weighted Class Activation Mapping for anatomical visualization.
  • ResNet-50 achieved the highest classification accuracy of 93.7%.
  • Macro-AUC was 0.97 in the class-balanced setting.
  • EfficientNetB3 showed superior generalization with fewer parameters.
  • Signal fidelity was preserved, ensuring critical diagnostic features were maintained.

Cite This Study

Neemuchwala et al. (2026) studied this question.

synapsesocial.com/papers/69aa701a531e4c4a9ff59963https://doi.org/10.3390/jimaging12030108
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Generalisation challenges in deep learning models for medical imagery: insights from external validation of COVID-19 classifiers2024 · 19 citations
  2. 2Beyond Post hoc Explanations: A Comprehensive Framework for Accountable AI in Medical Imaging Through Transparency, Interpretability, and Explainability2025 · 66 citations
  3. 3Image biomarkers and explainable AI: handcrafted features versus deep learned features2024 · 90 citations
  4. 4Explainability and uncertainty: Two sides of the same coin for enhancing the interpretability of deep learning models in healthcare2025 · 66 citations
  5. 5A Comprehensive Survey of Deep Learning Approaches in Image Processing2025 · 123 citations