FasteNet: A Fast Railway Fastener Detector

In this work, a novel high-speed railway fastener detector is introduced. This fully convolutional network, dubbed FasteNet, foregoes the notion of bounding boxes and performs detection directly on predicted saliency map. FasteNet uses transposed convolutions skips connections, effective receptive field network 1.5\(\times \) larger than average size fastener, enabling to make predictions with high confidence, without sacrificing output resolution. addition, due map approach, able vote for presence up 30 times per boosting prediction accuracy. capable running at 110 FPS an Nvidia GTX 1080, while taking in inputs 1600 \(\times 512 14 fasteners image. Our source open here: https://github.com/jjshoots/DL_FasteNet.git.