:zap: Yolo universal target detection model combined with EfficientNet-lite, the calculation amount is only 230Mflops(0.23Bflops), and the model size is 1.3MB
Code for our IEEE TIP 2020 paper "Dynamic Feature Integration for Simultaneous Detection of Salient Object, Edge and Skeleton"
List of useful data augmentation resources. You will find here some not common techniques, libraries, links to github repos, papers and others.
