I. Introduction
Convolutional Neural Networks (CNNs) have revolutionized computer vision and robotic perception with the introduction of neural network architecture models [1, 2, 3] that have consistently outperformed their non-CNN competitors since 2012 [4]. Significant advances in computational processing, primarily GPUs, are beginning to permit their usage for real-time autonomous MAV (Micro Aerial Vehicle) navigation where low run times are a necessity and additional heavy on-board sensors and processing units consume precious resources.