What question did this study set out to answer?

This research aims to assess how different input types affect the performance of deep learning models in automatic modulation classification.

June 11, 2026The Journal of Korean Institute of Electromagnetic Engineering and ScienceOpen Access

Performance Comparison of Deep Learning Based Automatic Modulation Classifiers According to Input Types

View Full Paper

Ask AI

Bookmark

Authors

SKSolsong Kang BSBo-Seok Seo

Discussion

Member takes

Overview

Randomized trial compares classification accuracy of deep learning models for automatic modulation classification, indicating input type significance.

Key Points

This research aims to assess how different input types affect the performance of deep learning models in automatic modulation classification.
A dataset of 26 modulation schemes was created for evaluation.
Five input types were tested: time-domain inphase/quadrature, time-domain magnitude/phase, frequency-domain real/imaginary, frequency-domain magnitude/phase, and time-frequency spectrogram.
Five deep learning models were employed, including CNN, LSTM, ResNet, Transformer, and hybrid CNN-LSTM, under varying signal-to-noise ratio conditions.
LSTM model using time-domain inphase and quadrature inputs achieved the highest classification accuracy with a significant margin.
Time-domain inputs generally provided better performance compared to frequency-domain inputs across models.
No notable difference in performance was found between the two time-domain input types.

Cite This Study

Kang et al. (2026) studied this question.

synapsesocial.com/papers/6a2a520a80c8f91e7f39e210 https://doi.org/10.5515/kjkiees.2026.37.5.492

View Full Paper

Ask AI

Bookmark

June 11, 2026The Journal of Korean Institute of Electromagnetic Engineering and ScienceOpen Access

Performance Comparison of Deep Learning Based Automatic Modulation Classifiers According to Input Types

View Full Paper

Ask AI

Bookmark

Authors

SKSolsong Kang BSBo-Seok Seo

Discussion

Member takes

Overview

Randomized trial compares classification accuracy of deep learning models for automatic modulation classification, indicating input type significance.

Key Points

This research aims to assess how different input types affect the performance of deep learning models in automatic modulation classification.
A dataset of 26 modulation schemes was created for evaluation.
Five input types were tested: time-domain inphase/quadrature, time-domain magnitude/phase, frequency-domain real/imaginary, frequency-domain magnitude/phase, and time-frequency spectrogram.
Five deep learning models were employed, including CNN, LSTM, ResNet, Transformer, and hybrid CNN-LSTM, under varying signal-to-noise ratio conditions.
LSTM model using time-domain inphase and quadrature inputs achieved the highest classification accuracy with a significant margin.
Time-domain inputs generally provided better performance compared to frequency-domain inputs across models.
No notable difference in performance was found between the two time-domain input types.

Cite This Study

Kang et al. (2026) studied this question.

synapsesocial.com/papers/6a2a520a80c8f91e7f39e210 https://doi.org/10.5515/kjkiees.2026.37.5.492

Also Consider

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

View Full Paper

Ask AI

Bookmark