Audio-to-Image Encoding for Improved Voice Characteristic Detection Using Deep Convolutional Neural Networks