HIGHLIGHTS
SUMMARY
Because of the influence of intricate factors, such as dramatic scale variance and complicated backgrounds filled with distractors, small objects (usually defined as objects with pixels below 32 × 32) in remote sensing images are always missed. In recent years, convolutional neural_networks (CNNs) have achieved significant progress in object detection tasks owing to their superior feature representation capabilities and the research in remote sensing images is currently dominated by them. Consequently, the authors modified YOLOv5 based on the Swin Transformer and propose a one-stage detection architecture for small-object detection in remote sensing images . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.