The Importance of the Instantaneous Phase for classification using Convolutional Neural Networks

TL;DR

This paper demonstrates that using the phase component of AM-FM representations as input for CNNs significantly improves classification efficiency and speed, achieving comparable performance with substantially fewer resources.

Contribution

The study introduces the use of AM-FM phase components as input features for CNNs, showing they outperform traditional grayscale images and reduce computational costs.

Findings

Phase component yields significant predictions in CNNs.

FM-based approach is 7x faster in training.

Uses 123x fewer parameters than MobileNetV2 with similar accuracy.

Abstract

Large-scale training of Convolutional Neural Networks (CNN) is extremely demanding in terms of computational resources. Also, for specific applications, the standard use of transfer learning also tends to require far more resources than what may be needed. This work examines the impact of using AM-FM representations as input images for CNN classification applications. A comparison was made between AM-FM components combinations and grayscale images as inputs for reduced and complete networks. The results showed that only the phase component produced significant predictions within a simple network. Neither IA or gray scale image were able to induce any learning in the system. Furthermore, the FM results were 7x faster during training and used 123x less parameters compared to state-of-the-art MobileNetV2 architecture, while maintaining comparable performance (AUC of 0.78 vs 0.79).