Using Generative Adversarial Networks to Augment Unmanned Aerial Vehicle Image ClassifixC;cation Training Sets

A challenging task in computer vision is fixC;nding techniques to improve the object detection and classixC;fication capabilities of ML models used for processing images acquired by moving aerial platforms. This research explores if GAN augmented UAV training sets can increase the generalizability of a detection model trained on said data. To answer this question, the YOLOv4-Tiny Object Detection Model was trained with aerial image training sets depicting rural environments. The salient objects within the frames were recreated using various GAN architectures, placed back into the original frames, and the augmented frames appended to the original training sets. GAN augmentation on aerial image training sets led to a 6.75 increase on average in the mAP of the YOLOv4-Tiny Object Detection model with a best-case increase of 15.76 . Similarly, a 4.13 increase on average and a best-case increase of 9.60 was observed for the IoU rate. Finally, 100.00 TP, 4.70 FP, and zero FN detection rates were yielded, providing further evidence supporting GAN augmentation for object detection model training sets.